Coke discharging process of semi-coke

The invention relates to a novel coke discharging process of semi-coke. According to the coke discharging process of the semi-coke, two coke discharging valves, namely an upper coke discharging valve and a lower coke discharging valve are utilized, wherein each coke discharging valve has a gas locking function, a cooling type coke extractor is arranged between the two coke discharging valves, and the cooling type coke extractor is provided with a water-cooled jacket and a nitrogen gas seal. The coke discharging valve on the upper portion of the cooling type coke extractor is a multi-sectional lattice type spiral coke discharging valve where coke can be discharged continuously and uniformly, the coke discharging valve on the lower portion of the cooling type coke extractor is a lattice type material feeding valve, and a chain type conveyor below the cooling type coke extractor is provided with an enclosed type shell structure. The system of the coke discharging process of the semi-coke is ...

Radio-frequency starting method of medical rotary accelerator

The invention belongs to the technical field of accelerators and discloses a radio-frequency starting method of a medical rotary accelerator. The method comprises the following steps of: adopting radio-frequency pulse for starting, wherein radio-frequency signals adopt cold-state resonance frequency of a cavity body as the center; carrying out circular frequency scanning in the range of +/-0.2MHz and under the conditions that the stepping is 0.001MHz and the interval is 0.5s; monitoring the sizes of reflecting signals and cavity-body sampling signals in the process of circular frequency scanning of the pulse, and directly switching into a continuous state by an analog switch; then tuning the frequency change between a hot state and the cold state of the cavity body by a PLL (Phase Locked Loop); when the cavity body is close to heat balance, starting a fine-tuning capacitance board; and finally increasing the power to beam-ejecting power after stabilizing the radio-frequency system to the ...

Direct torque control method and device of brushless continuous current dynamo

The invention belongs to the field of control technology of brushless continuous current dynamo, and discloses a direct torque control method of a brushless continuous current dynamo. The method comprises the following steps: A. a preset torque is obtained according to the actual rotational speed and the command rotational speed of the dynamo; B. an estimating torque is obtained according to the rotor position and the three-phase current value of the input end of the dynamo; C. a first duty ratio is obtained according to the preset torque and the estimating torque, and is added with a second duty ratio related to the speed, and then a final duty ratio is obtained; D. the state of a switch tube is selected according to the final duty ratio, the rotor position and the plus-minus of the preset torque; E. an inverter is controlled to drive the dynamo to rotate according to the state of the switch tube. The invention also discloses a direct torque control device of the brushless continuous current ...

Vertical symmetrical welding method for wide and thick plates

The invention discloses a vertical symmetrical welding method for wide and thick plates. The method is capable of guaranteeing planeness of vertical plates not more than 0.2mm, and includes the steps of a, welding to form a left cutter head, a middle plate and a right cutter head; b, heat treatment; c, finish machining; and d, welding of the left cutter head, the middle plate and the right cutter head by means of vertical up symmetric gas metal arc welding. For the vertical up symmetric gas metal arc welding, argon added with carbon dioxide is used as protective gas, wherein mass content of the argon is 80%, and mass content of the carbon dioxide is 20%; and a first composite weld and a second composite weld are identical in groove dimension, wherein groove width satisfies 18mm Подробнее

Middle and high energy neutron detector

The invention relates to a middle and high energy neutron detector. The detector comprises an outer slowing down body (2) arranged on an outermost layer, an absorbed layer (4) arranged in the outer slowing down body (2) and an inner slowing down body (5) arranged in the absorbed layer (4). An attenuation layer (3) is arranged between the outer slowing down body (2) and the absorbed layer (4). The outer slowing down body (2), the attenuation layer (3), the absorbed layer (4) and the inner slowing down body (5) comprise a common center hole (6). A low energy neutron probe (11) is arranged in the center hole (6). A leading-out wire of the low energy neutron probe (11) is connected with a counter arranged on an external portion of the middle and high energy neutron detector. The outer slowing down body (2) and the inner slowing down body (5) are made of polyethylene. The attenuation layer (3) is made of pure lead. The absorbed layer (4) is made of boracic polythene. The detector possesses the ...

Three-dimensional well pattern model for in-situ mining of shale oil

The invention relates to a three-dimensional well pattern model for in-situ mining of shale oil. The three-dimensional well pattern model integrally consists of a reaction kettle body and a rotary inclining mechanism, wherein the reaction kettle body is arranged above the rotary inclining mechanism. The reaction kettle body sequentially comprises a box wall, a heat preserving layer, a strong microwave absorption medium layer, a simulation well pattern and a temperature and pressure measuring point from outside to inside. According to the invention, the model not only can eliminate the microwave influence, adequately consider the dip angle situation under the effect of the gravity and can simulate a real stratum, but also can flexibly adjust the position and quantity of simulation wells to form different well pattern layouts. Due to the adoption of the model, the obtained experimental data is more accurate and scientific. Reliable technical parameter support can be provided for the development ...

Combustion device for processing calcined petroleum coke smoke of rotary kiln

The invention discloses a combustion device for processing calcined petroleum coke smoke, particularly relates to a combustion device for processing calcined petroleum coke smoke of a rotary kiln. The combustion device for processing the calcined petroleum coke smoke of the rotary kiln comprises a combustion chamber, a smoke inlet and a smoke outlet, wherein a dust collecting device is arranged at the lower end of the combustion chamber of vertical cylindrical shape. A fuel burner, an air supply duct and a combustion air fan are arranged on the combustion chamber, wherein the combustion air fan is connected with the air supply duct through a pipeline, and the air supply duct is installed on the middle and lower portion of the combustion chamber. A horizontal smoke combustion chamber in prior art is changed to a vertical cylindrical smoke combustion chamber, different components are reasonably arranged, and the structure is simple, so that the covering area of the device is greatly reduced ...

Method, device and system for remote startup and shutdown

The invention discloses a method for remote startup and shutdown, which is used for realizing remote startup and shutdown. The method comprises the following steps: receiving a remote control signal sent by a remote control device, and demodulating the remote control signal to be a digital signal; identifying a startup instruction or a shutdown instruction from the digital signal; and executing a computer startup operation or a computer shutdown operation according to the startup instruction or the shutdown instruction. The invention further discloses a device and a system for realizing the method.

Lifting overturning device of furnace door repair lifting frame

The invention relates to the field of furnace doors of coking furnaces, in particular to a lifting overturning device of a furnace door repair lifting frame. The lifting overturning device is characterized in that the lifting overturning device is an electric and manual dual-purpose mechanism driven by a single motor, a first lifting screw and a second lifting screw are arranged on the inner side of a dual-track frame in parallel, a sliding box body is movably connected with stand columns on the two sides of the dual-track frame, the top end of the sliding box body is respectively connected with the first lifting screw and the second lifting screw through two sets of lifting worm wheels to form a worm wheel-screw lifting mechanism, a first gear is arranged on a synchronous worm shaft between the two sets of the lifting worm wheels, a second gear is arranged on an output shaft of the motor, the second gear is meshed with the first gear when the second gear is located on a lifting station ...

Carbon calcining rotary kiln

The invention discloses a rotary kiln in the field of aluminum-used anode production, and particularly relates to a carbon calcining rotary kiln which comprises a kiln body, a lining, a kiln head, a kiln tail, a secondary air supply system and a tertiary air supply system. The secondary air supply system comprises a secondary air fan, a front secondary air pipe and a rear secondary air pipe, the front secondary air pipe and the rear secondary air pipe comprise 3-6 air pipes respectively, the air pipes are circularly arranged on the kiln body equidistantly, the straight-line distance between each two adjacent air pipes is 1.2-1.5m, the tertiary air supply system is composed of a tertiary air fan, a front tertiary air pipe and a rear tertiary air pipe, the front tertiary air pipe and the rear tertiary air pipe are composed of 3-6 air pipes respectively, the air pipes are circularly arranged on the kiln body equidistantly, and the straight-line distance between each two adjacent air pipes ...

Liquefied natural gas (LNG) low temperature liquefied three-level refrigeration spiral wound tube type heat exchanger

The invention belongs to the field of natural gas low temperature liquefied technology, and relates to a liquefied natural gas (LNG) three-level low temperature liquefied process device and a mixed refrigerant refrigeration technology, which cools 6.0MPa, -120DEG C natural gas into -164DEG C to be liquefied in two-steam spiral wound tube type heat exchanger by applying a certain proportion of non-azeotropic N2-CH4 mixed refrigerant, thereby being convenient for LNG to be stored in a surpercooling mode and being convenient to transport. The non-azeotropic N2-CH4 mixed refrigerant is firstly precooled and liquefied by a three-level refrigerant spiral wound tube type heat exchanger, and is throttled to a shell pass after being liquefied to cool a LNG tube bundle and a N2-CH4 mixed refrigerant precooling tube bundle from a secondary outlet in -120DEG C, thereby leading natural gas and precoolded refrigerant in the tube pass to be liquefied, and achieving the purpose of precooling mixed refrigerant ...

Method for molding 90-degree elbow pipe after surface welding of inner wall of pipe billet of straight pipe

The invention discloses a method for molding a 90-degree elbow pipe after surface welding of an inner wall of a pipe billet of a straight pipe used in a large-scale pressure container. The method can shorten the production-manufacturing cycle and guarantee the quality of the surface welding layer efficiently. The method for molding the 90-degree elbow pipe after surface welding of the inner wall of the pipe billet of the straight pipe comprises the steps of: a, selecting the pipe billet of the straight pipe; b, surface-welding a transitional layer at the inner wall of the pipe billet of the straight pipe by adopting an automatic surface welding machine, wherein the transitional layer is formed by surface welding through argon tungsten-arc welding; c, surface-welding an anti-corrosion layer on the surface of the transitional layer by adopting the automatic surface welding machine, wherein the anti-corrosion layer which is formed by surface welding through argon tungsten-arc welding; d, heating ...

Switch magnetic resistance generator system capable of realizing rapid excitation/demagnetization

The invention discloses a switch magnetic resistance generator system capable of realizing rapid excitation/demagnetization. The system comprises a switch magnetic resistance motor body, a prime mover, a position sensor, a power converter, a DC power supply and a controller, wherein the power converter is a four-level power converter. By adopting the four-level power converter, at an excitation stage, in an angle position control mode, the exciter current of a phase winding is increased, and the energy storage of a magnetic field is increased so that the output power of an overall power generation system is improved and it is easier for the power generation system to reach a rated power; in an current chopping control mode, the excitation time is reduced, rapid excitation is realized, and the time of a generating stage is increased so that the generating capacity of a motor is enhanced and the output power of the system is increased; and at the generating system, when two voltage sources ...

Inserted low temperature condensing plate device

The invention relates to a low temperature condensing plate device. The low temperature condensing plate device comprises low temperature condensing plates (3), wherein a shielding plate (2) formed by connecting a lower bottom plate and a side panel is arranged around the low temperature condensing plates (3). A baffle (1) is arranged at the top of the shielding plate (2). The lower bottom plate of the shielding plate (2) is connected with a primary cold head of a refrigerator (6). Two sets of low temperature condensing plates (3) with holes are respectively connected with secondary cold heads of two double secondary refrigerators (6). The refrigerators (6) are fixed on an installation flange (4). A low temperature condensing device is arranged in a vacuum room through the installation flange (4). Therefore, the low temperature condensing plate device overcomes the problem that gas extraction speed is limited by the diameter of a pump opening of a low temperature pump and the baffle in ...

Die-cast zinc alloy with high flowability and preparation method thereof

The invention discloses die-cast zinc alloy with high flowability and a preparation method thereof. The alloy comprises the components by weight percent: 4.51-6.0% of Al, 0.01-0.08% of Mg, 0.5-0.7% of Cu, 0.01-0.02% of Li, 0-0.1% of Ti, 0-0.02% of B, less than 0.02% of inevitable impurity elements such as Pb, Cd, Fe, Sn in total, and the balance of Zn. The preparation method comprises the steps of: firstly, putting an Al ingot into a smelting furnace, heating to 680-720 DEG C; adding a covering agent after a part of molten aluminum appears; cooling to 650-680 DEG C after the Al ingot is completely melted; adding a Zn ingot, Al-Mg, Al-Cu and Zn-Li middle alloy, refining and drossing; adding B and Ti elements, agitating, drossing and cooling to 420-560 DEG C; and stewing, drossing and casting. Compared with common die-cast zinc alloy, the flowability of the die-cast zinc alloy provided by the invention can be improved by over 70%; the grain size of the die-cast zinc alloy is effectively refined ...

Spray-type evaporative cooling and circulating system of heating device

The invention provides a spray-type evaporative cooling and circulating system of a heating device. A plurality of element seal boxes (110) are arranged in a heating device cabinet (120), a control valve (20) is arranged between a liquid storage tank (10) and a circulating pump (30), a filter (40) is arranged on the front end of a liquid distributor (50), the liquid distributor (50) is arrange on any position in the heating device cabinet (120), a plurality of nozzles (60) are installed on the liquid distributor (50), and one end of a gas collecting tube (70) is connected with the element seal boxes (110). After an evaporative cooling medium atomized by the nozzles (60) exchanges heat with a heating unit (80), the liquid part in the evaporative cooling medium returns to the liquid storage tank (10) through a liquid return tube (100), the evaporated evaporative cooling medium rises to a condenser (90) through the gas collecting tube (70), is condensed by the condenser (90) to become liquid ...

Technology for removing and stopping algae for ship taking solar energy as power

The invention discloses a technology for removing and stopping algae for ship taking solar energy as power. The principle of the technology lies in that a certain amount of solar panels are arranged on the ship, solar energy is converted into electricity for storage and used as power energy of the ship and a power supply of equipment on the ship. Ultrasonic algae removal and stopping equipment and radio radiation field algae removal and stopping equipment are arranged on the ship. The purposes of enabling the algae plants to die or stopping the growth of the algae plants can be achieved by enabling the sweep frequency ultrasonic wave with specific frequency band meeting certain power requirement, and specific frequency band electromagnetic field to be acted in water area containing a great amount of algae plants and breaking the living environment of the algae plants.

Combined type electronic laser music fountain

The invention discloses a combined type electronic music laser fountain which comprises an atomization generation device and parallel connection laser rotary tables, wherein the atomization generation device converts water into water mist through an atomizer and outputs the water mist to a mist cover, and an atomization environment is formed in the mist cover; a sound device is mounted on the atomization generation device; a cavity used for mounting the parallel connection laser rotary tables is formed in the mist cover; each parallel connection laser rotary table comprises a motor driving connecting rod transmission mechanism, and the connecting rod transmission mechanisms drive two-freedom-degree parallel connection rotating mechanisms to move and laser emission components mounted at the top parts of the parallel connection rotating mechanisms to rotate, so that laser emission at different angles is realized; and laser can generate light beams in the atomization environment, and the light ...

Catalyst coating method for particle oxidation catalytic reactors

The invention discloses a catalyst coating method for particle oxidation catalytic reactors. The method comprises follows steps: first, washing the Fe-Cr-Al metal carrier to get rid of absorbed impurities and oxides on the surface of the metal carrier, subjecting the metal carrier to a high-temperature oxidation treatment in a temperature range of 900 DEG C to 1050 DEG C to make Al in the metal carrier diffuse to the surface of the carrier and become a Al2O3 thin film; secondly, coating a MgAl2O4 transition layer on the surface of the metal carrier; thirdly, adding zirconium acetate with a weight percentage of 10% to 23%, barium acetate with a weight percentage of 0.1 to 0.4%, and manganous nitrate with a weight percentage of 1% to 3% in the catalyst coating paste, homogenously mixing, then coating the mixture on the MgAl2O4 transition layer, and finally a catalyst coating is formed after the processes of drying and burning. The catalyst coating method for particle oxidation reactors is ...

Instantaneous torque control method of brushless direct-current motor based on direct-current control

The invention discloses an instantaneous torque control method of a brushless direct-current motor based on direct-current control. The control method comprises the following steps: a two-phase current value of the motor is detected by using a current sensor and a current value of another phase is calculated; a rotor position of the motor is detected by using a position sensor; the instantaneous torque required at present is obtained through calculation and comparison; an optimal reference current is obtained through a torque coefficient reciprocal database in combination with the rotor position; the optimal reference current is compared with a current obtained through the selection so as to obtain a duty ratio D1; the duty ratio D1 is added to a duty ratio D2 corresponding to a counter potential of which the rotate speed is in direct ratio so as to obtain a final duty ratio D; the final duty ratio D, a rotate speed instruction, the negative and positive of the optimal reference current ...

Impressed current negative pole protective device of self-elevating platform and protective method thereof

The invention relates to the technical field of metal corrosion prevention, in particular to an impressed current negative pole protective device of a self-elevating platform and a protective method thereof. The device comprises an auxiliary positive pole (4), a reference electrode (7), current control equipment (1) and a cable, wherein the current control equipment (1) is arranged on the self-elevating platform (9), the auxiliary positive pole (4) and the reference electrode (7) are fixed on a pile shoe (8) of the self-elevating platform (9) through additional plates (5), and the auxiliary positive pole (4) and the reference electrode (7) are electrically connected with the current control equipment (1) through the cable which can be coiled and uncoiled in a longitudinal mode. The cable enters into the interior of the pile shoe (8) of the self-elevating platform (9) through a water sealing part. The impressed current negative pole protective device is simple in structure, convenient to ...

Sodium dichlorophenolate self-emulsifying ointment

The invention relates to a sodium dichlorophenolate self-emulsifying ointment, which can be used for effectively reducing first pass effect and enhancing the curative effect. According to the technical scheme of the invention, the sodium dichlorophenolate self-emulsifying ointment is prepared by uniformly mixing sodium dichlorophenolate serving as a medicament and a self-emulsifying substrate serving as a medicament carrier; every gram of sodium dichlorophenolate self-emulsifying ointment contains 20-50mg of sodium dichlorophenolate; the self-emulsifying substrate is prepared from 12-20 percent by weight of an oil phase, 40-60 percent by weight of an emulsifier and 20-40 percent by weight of an auxiliary emulsifying agent; the oil phase consists of refined oil, glycerin monostearate and octadecanol in the weight ratio of 8:2:1; the emulsifier consists of polysorbate 80 and RH40 in the weight ratio of 2:8, or consists of polysorbate 80 and EL in the weight ratio of 6:4; and the auxiliary ...

Method and device for compressing flow data

A method for compressing flow data, including: constructing multiple line segments according to flow data and a predefined maximum error that are acquired; obtaining a target piecewise linear function according to the multiple line segments, where the target piecewise linear function includes multiple linear functions, and an intersection set of value ranges of independent variables of every two linear functions among the multiple linear functions includes a maximum of one value; and outputting a reference data point according to the target piecewise linear function, where the reference data point includes a point of continuity and a point of discontinuity of the target piecewise linear function. In this way, a maximum error, a target piecewise linear function is further determined according to the multiple line segments, and a point of continuity and a point of discontinuity of the target piecewise linear function are used to represent compressed flow data.

Method and device for processing audio/video data

The invention discloses a method and a device for processing audio/video data. The processing method comprises the following steps: acquiring audio data acquired through a microphone and video data acquired through a camera; encoding the acquired audio data and video data respectively to obtain encoded audio data and video data; processing the encoded audio data by adopting a self-adaptive echo cancellation algorithm to form processed audio data; encapsulating the encoded video data into a video data package and encapsulating the processed audio data into an audio data package according to the format of a real-time messaging protocol (RTMP); transmitting the video data package and the audio data package to a RTMP server. According to the technical scheme provided by the invention, audio and video delays can be reduced, and the quality of audio/video data is improved.

Ultrahigh molecular weight polyethylene composite material and manufacturing method thereof

The invention relates to an ultrahigh molecular weight polyethylene composite material which is characterized by being formed by multiple fiber layers superposed through an adhesive, wherein the fiber layers are independently layers formed by ultrahigh molecular weight polyethylene fiber or layers formed by reinforced fiber. The formula of the adhesive based on 100wt% of the adhesive is as follows: 5-30wt% of any one or more cementing compound resins of EVA (Ethylene-Vinlacetate Copolymer) glue, SBS (Styrene-Butadiene-Styrene) glue, SEBS (Styrene-Ethylene-Butylene-Styrene) glue, SIS (Styrene-Isoprene-Styrene) glue, TPR (Thermoplastic Rubber) glue styrene glue and buna, 2-10wt% of any one or more tackifying resins of petroleum resin, rosin resin and terpene resin, and 60-90wt% of a hydrocarbon solvent. According to the invention, the fiber layers are bonded through the adhesive, so that the deficiencies of a single high performance fiber material are overcome, the integral performance of ...

Multi-sectional grid type spiral coke discharging valve

The invention relates to a multi-sectional grid type spiral coke discharging valve of field of coking. The device comprises a plurality of grid type coke discharging teeth; each grid type coke discharging tooth comprises a coke discharging tooth and a separating plate; the coke discharging tooth, the separating plate and a coke discharging box housing form a closed unit capable of locking air and uniformly discharging materials; coke discharging channels are arranged among the coke discharging teeth; two adjacently arranged coke discharging teeth of two grid type coke discharging teeth are staggered in a certain angle; and the coke discharging tooth generating lines are distributed as spiral lines. The rotation speed of the multi-sectional grid type spiral coke discharging valve can be adjusted by adjusting the flow rate of hydraulic oil or the rotation speed of a motor according to the requirement of production technology of the semi coke. The multi-sectional grid type spiral coke discharging ...

Method for determining dissolved oxygen in wetland plant rhizosphere micro-interface in situ

The invention discloses a method for determining dissolved oxygen in a wetland plant rhizosphere micro-interface in situ, which comprises the following steps: a, providing a fixed carrier of a wetland plant root system; b, horizontally placing a wetland plant rhizosphere and fixing the wetland plant rhizosphere on the surface of the fixed carrier of the wetland plant root system; c, longitudinally displacing a dissolved oxygen microelectrode to a determining part of the wetland plant rhizosphere, then displacing the dissolved oxygen microelectrode in the reverse direction, recording the displacement distance, adding sediment to cover the wetland plant rhizosphere and cultivating a first predetermined time; and d, resetting the dissolved oxygen microelectrode to the determining part of the wetland plant rhizosphere and determining the concentration of dissolved oxygen on the surface of the determining part of the wetland plant rhizosphere. According to the method provided by the invention ...

Technology for water de-scaling and sterilizing by radio radiation field

The invention discloses a technology for water de-scaling and sterilizing by a radio radiation field. According to the technical theory, a radio antenna with omni-directional radiation is arranged at the centre line of a flow pipeline and radiates electromagnetic waves with certain power and variable frequency to the four sides of the pipeline so as to form the electromagnetic field acting on water passing through the pipeline, thereby fulfilling the purposes of water de-scaling and sterilizing. The radiation energy of the antenna is provided by a swept frequency signal source. The swept frequency signal source can generate swept frequency signals of which the frequencies are changed in two frequency bands (wherein one frequency band is used for de-scaling, and the other frequency band is used for sterilizing), and can realize continuous frequency sweeping, modulated frequency sweeping, and continuous and modulated mixed frequency sweeping. After being amplified to curtain power, the swept ...

Purifying method for black and odorous water and sediment nutrient salt

The invention discloses a purifying method for black and odorous water and sediment nutrient salt. The purifying method comprises the following steps: a) carrying out aeration treatment on the surface sediment and the overlying water of the black and odorous water to synchronously reduce the ammonia nitrogen contents of the water and the surface sediment, and meanwhile immobilizing the phosphorus in the water by adsorption; b) carrying out settling treatment on the water subjected to the aeration treatment; and c) repeatedly carrying out the treatments in the steps a) and b) on the water subjected to the settling treatment until the concentration of ammonia nitrogen in the water is lower than a predetermined concentration, keeping the water standing still, and finishing the purification of the water. According to the purifying method provided by the invention, the adding of a chemical agent is not needed, while all that is needed to do is change the position of an aerator on the basis of ...

Protective circuit for preventing radio station from uninterrupted transmission

The invention relates to a protective circuit for preventing a radio station from uninterrupted transmission, and belongs to the technical field of power communication. A power supply of the circuit and a connection circuit of a wireless data transmission radio station are connected in series with a switch control device; a current detection control circuit is branched by a sampling device; the current detection control circuit comprises a comparison device; the sampling signal output end of the sampling device is connected with the comparison signal input end of the comparison device; and the control signal output end of the comparison device is connected with the controlled end of the switch control device, and is used for outputting a control signal for switching off the switch control device when the input current detection signal is greater than a threshold value and exceeds a predetermined time length, so that the effect of preventing the radio station from uninterrupted transmission ...

METHOD AND DEVICE FOR COMPRESSING FLOW DATA

A method for compressing flow data, including: constructing multiple line segments according to flow data and a predefined maximum error that are acquired; obtaining a target piecewise linear function according to the multiple line segments, where the target piecewise linear function includes multiple linear functions, and an intersection set of value ranges of independent variables of every two linear functions among the multiple linear functions includes a maximum of one value; and outputting a reference data point according to the target piecewise linear function, where the reference data point includes a point of continuity and a point of discontinuity of the target piecewise linear function. In this way, a maximum error, a target piecewise linear function is further determined according to the multiple line segments, and a point of continuity and a point of discontinuity of the target piecewise linear function are used to represent compressed flow data. 1. A method for compressing flow data , comprising:acquiring flow data, wherein the flow data comprises multiple data points, and each data point of the multiple data points comprises collection time of each data point and a data value collected at the collection time;acquiring a maximum error that is predefined;constructing multiple line segments according to the multiple data points and the maximum error, wherein the multiple line segments do not intersect with each other, each line segment of the multiple line segments represents a function with time as an independent variable, and when a value of the independent variable is the collection time, an absolute value of a difference between a function value of the function and the data value collected at the collection time is less than or equal to the maximum error;obtaining a target piecewise linear function according to the multiple line segments, wherein the target piecewise linear function comprises multiple linear functions, and an intersection set of value ...

Coke oven heating method

The invention relates to the field of metallurgical coking, and discloses a coke oven heating method. The coke oven heating method is characterized in that air is supplied to a vertical heating flue of a combustion chamber in sectionalized mode, lean gas or rich gas or mixed gas is supplied from the bottom of the combustion chamber, the gas and the air are mixed and combust in sectionalized mode to generate waste gas, the generated waste gas rises to the top of the vertical heating flue, passes through crossing holes and turns to enter into a downdraft vertical heating flue, downdrafts passing through the bottom of the downdraft vertical heating flue pass through chutes and enter into an adjacent heat collecting chamber, a waste gas circulation hole is arranged in the root portion of a partition wall of two flues of a twinflue, the waste gas in the downdraft vertical heating flue is sucked into an updraft vertical heating flue through the waste gas circulation hole and mixed with the flow ...

Bead-welding method of inner wall of 90-degree bent pipe

The invention discloses a bead-welding method of an inner wall of a 90-degree bent pipe for a large-scale pressure vessel, which can shorten the manufacturing cycle and effectively ensure the bead-welding quality of a transition layer and an anti-corrosion layer. The method is characterized in that the method comprises the following steps: a. selecting a pipe billet of a straight pipe; b. heating the pipe billet of the straight pipe in a furnace, taking out the heated pipe billet of the straight pipe, and performing compression moulding to the pipe billet on a pressing machine by a mould to form the 90-degree bent pipe, wherein the ovality of the formed 90-degree bent pipe is controlled to be less than or equal to 3%; c. performing bead welding to the surface of the inner wall of the formed 90-degree bent pipe by an automatic bead-welding machine for the inner wall of the bent pipe to form the transition layer; and d. performing bead welding to the surface of the transition layer to form ...

Yixin xuezhikang capsule

The invention discloses a Yixin xuezhikang capsule, which comprises the following bulk drugs by weight: 400 parts of gen-seng, 1080 parts of fleece-flower root, 300 parts of pseudo-ginseng and 720 parts of haw; the method comprises the following steps: crushing gen-seng, radix pseudoginseng and haw, then heating by ethanol for a reflux extraction twice, wherein each time for 1 hour, merging an extract and filtering, and concentrating the filtrate to a concentrated solution with relative concentration of 1.1 - 1.2 for standby; filtering herb residue and merging with fleece-flower root, decocting for three times by adding water, wherein each time for 1 hour, merging the decocted liquid, concentrating to the relative concentration of 1.1 - 1.2 and merging with the ethanol concentrated solution, spray drying, uniformly mixing for encapsulating. The preparation of the invention has the functions of benefiting heart qi, promoting blood circulation and removing blood stasis, lowering lipid and ...

Novel soilless cultivation method of leeches

The invention relates to a novel soilless cultivation method of leeches, and belongs to the technical field of cultivation of aquatic animals. According to the method, top openings of net cages are closed, the net cages are fixed to fresh water holders, a certain distance exists between every two net cages, 100-500 artificially-bred seedlings of which the average weight is 10mg are thrown in each net cage of the base area being 0.5m<2>, after the average weight of the seedlings is greater than 500mg, and the seedlings are transferred into another net cages to continue being cultivated, or 20-200 wild seedlings of which the average weight is 0.5-4g are thrown in each net cage of the base area being 0.5m<2>; main food of the weight being 20 times that of the leeches is thrown in the cultivation process, the food is updated regularly, and the leeches are directly poured out to be harvested after being cultivated for 60-120 days. The average weight of the leeches exceeds 20g, and the size of ...

SYSTEMS AND METHODS FOR IDENTIFICATION OF FLUID AND SUBSTRATE COMPOSITION OR PHYSICO-CHEMICAL PROPERTIES

Techniques for identifying a composition of a target fluid using a set of vectors representing known residue patterns for a two or more fluids including said target fluid is provided. An exemplary method includes storing one or more digital measurements of residue for the target fluid, extracting one or more descriptive features from the measurements; and processing descriptive features to identify the composition of the target fluid. The processing includes using a machine learning algorithm trained with data linking residue morphology to fluid composition. A distance between a vector representing said one or more descriptive features and said set of vectors representing known residue patterns is determined, and a residue is assigned to one or more of the known residue patterns. 1. A method for identifying at least one of a physico-chemical property or composition of a target fluid using a set of vectors representing known residue patterns for a two or more fluids including said target fluid , comprising:acquiring one or more digital measurements of residue for said target fluid;extracting one or more descriptive features from at least one of said measurements; andprocessing one or more of said descriptive features to identify said at least one physico-chemical property or composition of said target fluid from said at least one of said measurements;wherein said processing includes using a matching algorithm or a machine learning algorithm trained with data linking residue patterns to fluid composition or physico-chemical properties.2. The method of claim 1 , further comprising:determining a distance between a vector representing said one or more descriptive features and said set of vectors representing known residue patterns, andassigning a residue to one or more of the known residue patterns that minimizes said distance.3. The method of claim 1 , wherein acquiring one or more digital measurements comprises capturing one or more images.4. The method of claim 1 , ...

Extract for Preventing of Treating Thrombotic Diseases

An extract for preventing or treating thrombotic diseases, particularly, an extract of at least one of leeches and earthworms having a molecular weight of no more than 5,800 daltons is provided, wherein the extract includes 15% to 38% amino acid, 40% to 60% saccharide and 0.3% to 1% hypoxanthine. Processes for preparation, pharmaceutical compositions and uses thereof are also provided. Compared to conventional arts, the extract has safety greatly improved and drug actions maintained and even improved.

META-LEARNING FOR CAMERA ADAPTIVE COLOR CONSTANCY

A processing entity generates a model for estimating scene illumination colour for a source image captured by a camera The processing entity acquires a set of images, captured by a respective camera, the set of images as a whole including images captured by multiple cameras; forms a set of tasks by assigning each image of the images set to a respective task such that images in the same task have in common that a the images are in a predetermined range; trains model parameters by repeatedly: selecting at least one of the tasks, forming an interim set of model parameters based on a first subset of the images of that task, estimating the quality of the interim set of model parameters against a second subset of the images of that task and updating the parameters of the model based on the interim set of parameters and the estimated quality. 1. A processing entity , the processing entity configured to:generate a model for estimating scene illumination colour for a source image captured by a camera the model generation comprising:acquiring a set of images, each image of the set of images having been captured by a respective camera, the set of images as a whole comprising images captured by multiple cameras;forming a set of tasks by assigning each image of the set of images to a respective task such that images in a same task have in common that a property of the images lies in a predetermined range; and selecting at least one of the tasks,', 'forming an interim set of model parameters in dependence on a first subset of the images of that task,', 'estimating a quality of the interim set of model parameters against a second subset of the images of that task, and', 'updating the parameters of the model in dependence on the interim set of parameters and the estimated quality., 'training parameters of the model by repeatedly2. The processing entity as claimed in claim 1 , wherein the predetermined range is defined such that the predetermined range corresponds to a set of colour ...

Image classification method, neural network training method, and apparatus

An image classification method, a neural network training method, and an apparatus are provided, and relate to the field of artificial intelligence, and specifically, to the field of computer vision. The image classification method includes: obtaining a to-be-processed image; and obtaining a classification result of the to-be-processed image based on a pre-trained neural network model, where the classification result includes a class or a superclass to which the to-be-processed image belongs. When the neural network model is trained, not only labels of a plurality of training images but also class hierarchy information of the plurality of training images is used. That is, more abundant information of the training images is used. Therefore, images can be better classified.

IMAGE CLASSIFICATION METHOD AND APPARATUS, AND IMAGE CLASSIFICATION MODEL TRAINING METHOD AND APPARATUS

This application relates to an image recognition technology in the artificial intelligence field, and provides an image classification method and apparatus, and an image classification model training method and apparatus. This application relates to the artificial intelligence field, and more specifically, to the computer vision field. The method includes: obtaining a to-be-processed image; and classifying the to-be-processed image based on a preset global class feature, to obtain a classification result of the to-be-processed image. The global class feature includes a plurality of class features obtained through training based on a plurality of training images in a training set. The plurality of class features in the global class feature are used to indicate visual features of all classes in the training set. 1. An image classification method comprising:obtaining a to-be-processed image; and wherein the global class feature comprises a plurality of class features obtained through training based on a plurality of training images in a training set,', 'wherein the plurality of class features in the global class feature are used to indicate visual features of all classes in the training set,', 'wherein all the classes in the training set are classes to which all the training images in the training set belong, and', 'wherein the training set comprises images in a base class and images in a novel class., 'classifying the to-be-processed image, based on a preset global class feature, to obtain a classification result of the to-be-processed image,'}2. The method according to claim 1 , wherein classifying the to-be-processed image claim 1 , based on the preset global class feature claim 1 , to obtain the classification result of the to-be-processed image comprises:extracting a feature vector of the to-be-processed image; 'wherein the candidate class is one or more of the plurality of classes indicated by the global class feature; and', 'determining, based on the feature ...

DATA PROCESSING METHOD AND APPARATUS

The method includes: obtaining a plurality of pieces of feature data; automatically performing two different types of nonlinear combination processing operations on the plurality of pieces of feature data to obtain two groups of processed data, where the two groups of processed data include a group of higher-order data and a group of lower-order data, the higher-order data is related to a nonlinear combination of m pieces of feature data in the plurality of pieces of feature data, and the lower-order data is related to a nonlinear combination of n pieces of feature data in the plurality of pieces of feature data, where m≥3, and m>n≥2; and determining prediction data based on a plurality of pieces of target data, where the plurality of pieces of target data include the two groups of processed data. 1. A data processing method , comprising:obtaining a plurality of pieces of feature data;automatically performing two different types of nonlinear combination processing operations on the plurality of pieces of feature data to obtain two groups of processed data, wherein the two groups of processed data comprise a group of higher-order data and a group of lower-order data, wherein the higher-order data is related to a nonlinear combination of m pieces of feature data in the plurality of pieces of feature data, and wherein the lower-order data is related to a nonlinear combination of n pieces of feature data in the plurality of pieces of feature data, wherein m≥3, and m>n≥2; anddetermining prediction data based on a plurality of pieces of target data, wherein the plurality of pieces of target data comprise the two groups of processed data.2. The method according to claim 1 , wherein automatically performing the two different types of nonlinear combination processing operations on the plurality of pieces of feature data to obtain two groups of processed data comprises:automatically performing a higher-order nonlinear combination processing operation on the plurality of ...

Neural Network Search Method, Apparatus, And Device

A neural network search method, apparatus, and device are provided, and relate to the field of artificial intelligence technologies, and specifically, to the field of automatic machine learning technologies. The method includes: A computing device obtains a dataset and N neural networks (S602), where N is a positive integer; and performs K evolutions on the N neural networks to obtain neural network obtained through the Kth evolution, where K is a positive integer (S604). In a process of each evolution, a network structure of a neural network obtained in previous evolution is mutated; and a candidate neural network is selected, based on a partially ordered hypothesis, from networks obtained through mutation.

Data Processing Method and Apparatus

A data processing method, includes receiving a data flow; generating a triplet set according to the data flow, where each triplet in the set includes three items, the first item is a first element in the data flow, the second item includes a first time point at which the first element appears in the data flow and a first quantity of times that corresponds to the first time point, and the third item includes a second time point at which the first element appears in the data flow and a second quantity of times that corresponds to the second time point; and performing data processing on the data flow according to the triplet set. In the embodiments of the present application, the triplet set may be generated based on the data flow.

METHOD AND APPARATUS FOR PROCESSING DEVELOPMENT MACHINE OPERATION TASK, DEVICE AND STORAGE MEDIUM

The present application discloses a method and an apparatus for processing a development machine operation task, a device and a storage medium, which relates to the field of deep learning of artificial intelligence. The specific implementation solution is: receiving a task creating request initiated by a client; generating, according to the task creating request, the development machine operation task; allocating a target graphics processing unit (GPU) required for executing the development machine operation task for the development machine operation task; and sending a development machine operation task request to a master node in cluster nodes, where the task request is used to request executing the development machine operation task on the target GPU. 1. A method for processing a development machine operation task , comprising:receiving a task creating request initiated by a client;generating, according to the task creating request, a development machine operation task;allocating a target graphics processing unit (GPU) required for executing the development machine operation task for the development machine operation task; andsending a development machine operation task request to a master node in cluster nodes, wherein the task request is used to request executing the development machine operation task on the target GPU.2. The method according to claim 1 , wherein the allocating a GPU required for executing the development machine operation task to the development machine operation task comprises:determining a user group to which the development machine operation task belongs, wherein different user groups correspond to different resource usage rights; andallocating, according to a resource usage right corresponding to the user group to which the development machine operation task belongs and resources required for the development machine operation task, the target GPU required for executing the development machine operation task.3. The method according to claim ...

Self-Driving Method, Training Method, and Related Apparatus

A self-driving method and a related apparatus, the method including determining, by a self-driving apparatus, a task feature vector of a self-driving task according to M groups of historical paths of the self-driving task, where the task feature vector is a vector representing features of the self-driving task, and where M is an integer greater than 0, determining, by the self-driving apparatus, according to the task feature vector and a status vector, a target driving operation that needs to be performed, where the status vector indicates a driving status of the self-driving apparatus, and performing, by the self-driving apparatus, the target driving operation.

MUTANT P-HYDROXYPHENYLPYRUVATE DIOXYGENASE, AND CODING NUCLEIC ACID AND USE THEREOF

The present invention relates to a mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein or a bioactive fragment thereof and an isolated polynucleotide comprising a nucleic acid sequence encoding the protein or fragment thereof, wherein the mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein or a bioactive fragment thereof retains or enhances the property of catalyzing the conversion of p-hydroxyphenylpyruvate (HPP) to homogentisate and is significantly less sensitive to HPPD-inhibiting herbicides than a wild-type HPPD. The present invention also relates to a nucleic acid construct, an expression vector and a host cell comprising the polynucleotide, as well as to a method for producing a plant that has the property of catalyzing the conversion of p-hydroxyphenylpyruvate (HPP) to homogentisate and significantly reduced sensitivity to HPPD-inhibiting herbicides. 1. A mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein or a bioactive fragment thereof , wherein , as compared to the amino acid sequence of a wild-type rice p-hydroxyphenylpyruvate dioxygenase protein , the amino acid sequence of the mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein has one or more of mutations selected from the group consisting of 93S , 103S , 141R , 141K , 141T , 165V , 191I , 220K , 226H , 276W , 277N , 336D , 337A , 338D , 338S , 338Y , 342D , 346C , 346D , 346H , 346S , 346Y , 370N , 377C , 386T , 390I , 392L , 403G , 410I , 418P , 419F , 419L , 419V , 420S , 420T , 430G , and 431L at one or more positions corresponding to 93 , 103 , 141 , 165 , 191 , 220 , 226 , 276 , 277 , 336 , 337 , 338 , 342 , 346 , 370 , 377 , 386 , 390 , 392 , 403 , 410 , 418 , 419 , 420 , 430 and 431 of the amino acid sequence of wild-type rice p-hydroxyphenylpyruvate dioxygenase protein as set forth in SEQ ID NO: 2.2. The mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein or a bioactive fragment thereof according to claim 1 , wherein the amino acid sequence of the mutant p- ...

OLIGOSACCHARIDE COMPOUND FOR INHIBITING INTRINSIC COAGULATION FACTOR X-ENZYME COMPLEX, AND PREPARATION METHOD THEREFOR AND USES THEREOF

A purified oligosaccharide compound having antithrombotic activity or a mixture of a homologous compound thereof and a pharmaceutically acceptable salt thereof, a preparation method for the mixture, a pharmaceutical composition containing the mixture, and uses thereof serving as an intrinsic factor X-enzyme (Xase) inhibitor in the preparation of drugs for preventing and/or treating thrombotic diseases. 5. The oligosaccharide compound or a pharmaceutically acceptable salt thereof according to claim 1 , characterized in that claim 1 , in the general structure represented by Formula (I) claim 1 , n is optionally 1 claim 1 , 2 claim 1 , 3 or 4.6. The oligosaccharide compound or a pharmaceutically acceptable salt thereof according to claim 1 , characterized in that claim 1 , the pharmaceutically acceptable salt is optionally an alkali metal salt claim 1 , an alkaline earth metal salt or an organic ammonium salt.7. The oligosaccharide compound or a pharmaceutically acceptable salt thereof according to claim 6 , characterized in that claim 6 , the pharmaceutically acceptable salt is optionally a sodium salt claim 6 , a potassium salt or a calcium salt.8. An oligosaccharide mixture or a pharmaceutically acceptable salt thereof composed of the oligosaccharide compound of and having antithrombotic activity claim 6 , characterized in that claim 6 , the oligosaccharide mixture is composed of homologues of the oligosaccharide compound of ; and in the oligosaccharide compound of Formula (I) composing the oligosaccharide mixture claim 6 , Ris a group represented by Formula (II) claim 6 , Formula (III) claim 6 , or Formula (IV) claim 6 , and in the molar ratio claim 6 , the proportion of the oligosaccharide compound of Formula (I) in which Ris a group simultaneously represented by Formula (II) claim 6 , simultaneously represented by Formula (III) or simultaneously represented by Formula (IV) is not less than 95% in the mixture.9. A preparation method of the oligosaccharide compound ...

PREDICTION METHOD, TERMINAL, AND SERVER

Example prediction methods and apparatus are described. One example includes sending a first model parameter and a second model parameter by a server to a plurality of terminals. The first model parameter and the second model parameter are adapted to a prediction model of the terminal. The server receives a first prediction loss sent by at least one of the plurality of terminals. A first prediction loss sent by each of the at least one terminal is calculated by the terminal based on the prediction model that uses the first model parameter and the second model parameter. The server updates the first model parameter based on the first prediction loss sent by the at least one terminal to obtain an updated first model parameter. The server updates the second model parameter based on the first prediction loss sent by the at least one terminal to obtain an updated second model parameter. 1. A machine learning model training method , comprising:receiving, by a terminal, a first parameter and a second parameter of an algorithm model on a server that are sent by the server, wherein the first parameter is used as an initial parameter of a machine learning model of the terminal, and wherein the second parameter is used to indicate a manner in which the terminal trains the machine learning model;training, by the terminal, the machine learning model on the terminal based on the first parameter, the second parameter, and training set data, wherein the training set data comprises a plurality of training samples, wherein each training sample comprises a first eigenvector and a corresponding first label, and wherein the first eigenvector is constructed based on user information in historical data of the terminal;obtaining, by the terminal, a test result based on test data and the machine learning model;calculating, by the terminal, a loss gradient based on the test result; anduploading, by the terminal, the loss gradient to the server, wherein the loss gradient is used by the server ...

Graph partitioning method and apparatus

The method of the present disclosure includes: after a graph partitioning apparatus extracts an edge, first determining whether an aggregation degree between a currently extracted edge and an allocated edge in a first device satisfies a preset condition; then, when the preset condition is satisfied, determining whether a quantity of allocated edges stored in the first device is less than a first preset threshold; and allocating the currently extracted edge to the first device when the quantity is less than the first preset threshold. In this way, an aggregation degree between allocated edges in each device is relatively high and each device has relatively balanced load. When an edge changes and an edge associated with the particular edge needs to be synchronized, a relatively small quantity of devices need to perform synchronization and update, so that costs of communication between devices are reduced, and distributed graph computing efficiency is improved.

Recommendation Result Generation Method and Apparatus

A recommendation result generation method, where the method includes obtaining article content information of at least one article and user score information of at least one user, where user score information of a first user of the at least one user includes a historical score of the first user for the at least one article, encoding the article content information and the user score information using an article neural network and a user neural network respectively to obtain a target article latent vector of each of the at least one article and a target user latent vector of each of the at least one user, and calculating a recommendation result for each user according to the article latent vector and the user latent vector.

Machine Learning Model Training Method And Apparatus

Embodiments of the present invention provide a machine learning model training method, including: obtaining target task training data and N categories of support task training data; inputting the target task training data and the N categories of support task training data into a memory model to obtain target task training feature data and N categories of support task training feature data; training the target task model based on the target task training feature data and obtaining a first loss of the target task model, and separately training respectively corresponding support task models based on the N categories of support task training feature data and obtaining respective second losses of the N support task models; and updating the memory model, the target task model, and the N support task models based on the first loss and the respective second losses of the N support task models. 1. A machine learning model training method , comprising:obtaining target task training data and N categories of support task training data, wherein both the target task training data and the N categories of support task training data are labeled data, the target task training data corresponds to a target task model, the N categories of support task training data are in a one-to-one correspondence with N support task models, and N is a positive integer;inputting the target task training data and the N categories of support task training data into a memory model to obtain target task training feature data and N categories of support task training feature data, wherein both the target task training feature data and the N categories of support task training feature data are labeled data, the target task training feature data corresponds to the target task training data, and the N categories of support task training feature data are in a one-to-one correspondence with the N categories of support task training data;training the target task model based on the target task training feature ...

Similarity measurement method and device

A similarity measurement method includes: obtaining a directional relationship between nodes in a network, and determining a transition matrix; calculating a constraint matrix according to the transition matrix and an obtained attenuation factor; constructing a system of linear equations, where a coefficient matrix of the system of linear equations is the constraint matrix, and a variable of the system of linear equations is a correction vector; solving the system of linear equations by means of iteration by using a Jacobi method, and determining the correction vector; and calculating similarities between the nodes according to the transition matrix, the attenuation factor, and a diagonal correction matrix that is generated according to the correction vector. In the method, the correction vector is determined by using the Jacobi method, and further the similarities between the nodes may be calculated.

METHOD AND DEVICE FOR COMPRESSING FLOW DATA

A method for compressing flow data, including: generating multiple line segments according to flow data and a predefined maximum error that are acquired; obtaining a target piecewise linear function according to the multiple line segments, where the target piecewise linear function includes multiple linear functions, and an intersection set of value ranges of independent variables of every two linear functions among the multiple linear functions includes a maximum of one value; and outputting a reference data point according to the target piecewise linear function, where the reference data point includes a point of continuity and a point of discontinuity of the target piecewise linear function. In this way, a maximum error, a target piecewise linear function is further determined according to the multiple line segments, and a point of continuity and a point of discontinuity of the target piecewise linear function are used to represent compressed flow data. 1. A method for compressing flow data , comprising:acquiring, by a receiver, flow data comprising multiple data points over a period of time, each data point comprising a collection time and a data value collected at the collection time;generating, by a processor, multiple line segments in a two dimensional space over time and data value according to the multiple data points and a predefined maximum error, each line segment representing a function of time, wherein an absolute value of a difference between a function value of the function at the collection time and the data value collected at the collection time is no greater than the predefined maximum error;generating, by the processor, a representation of a target piecewise linear function defined over the period of time according to the multiple line segments, wherein the target piecewise linear function comprises multiple linear functions separately defined over a sequence of time segments along the period of time, wherein two adjacent time segments of the ...

Oil-gas separator and engine

本实用新型涉及发动机技术领域，尤其涉及一种油气分离器及发动机。油气分离器包括：壳体，其包括上盖和缸体，所述上盖和所述缸体连接；滤芯，其位于所述壳体的腔室内，所述滤芯的两端分别与所述上盖和所述缸体螺纹连接。滤芯的顶部和底部即两端分别与上盖和缸体螺纹连接，实现滤芯两端的固定，当发动机震动时，滤芯两端被固定则不会出现滤芯自身振动幅度增大的问题，油液不会在滤芯振动的时候被甩出，从而提高了油气分离效率。当发动机振动时，油气分离器的振动幅度小，油气分离效果好。

Neural network training method, image classification system, and related device

Embodiments of this application disclose a neural network training method, an image classification system, and a related device, which may be applied to the artificial intelligence field. Feature extraction is performed on images in a training set (including a first set and a second set) by using a prototype network, to obtain first feature points, in a feature space, of a plurality of images in the first set and second feature points of a plurality of images in the second set. The first feature points are used for calculating a prototype of a class of an image, and the second feature points are used for updating a network parameter of the prototype network. A semantic similarity between classes of the images in the second set is obtained, to calculate a margin value between the classes of the images. Then, a loss function is adjusted based on the margin value. The adaptive margin value is introduced into the loss function. To be specific, a larger margin value is added between similar classes, and a smaller margin value is added between less similar classes, to adaptively adjust a margin between classes, so as to more easily distinguish between classes, thereby improving classification accuracy of the prototype network.

Method and apparatus for processing development machine operation task, device and storage medium

The present application discloses a method and an apparatus for processing a development machine operation task, a device and a storage medium, and relates to the field of deep learning of artificial intelligence. The specific implementation solution is: receiving a task creating request initiated by a client; generating, according to the task creating request, the development machine operation task; allocating a target graphics processing unit (GPU) required for executing the development machine operation task for the development machine operation task; and sending a development machine operation task request to a master node in cluster nodes, where the task request is used to request executing the development machine operation task on the target GPU. Compared with the prior art, the present application can directly use an operating system of a local host by using a docker container to execute the development machine operation task on the GPU, thereby improving the utilization rate of the hardware of a physical machine.

Prediction method, and terminal and server

Embodiments of the present invention disclose a prediction method, a terminal, and a server. The method includes: sending, by a server, a first model parameter and a second model parameter to a plurality of terminals, where the first model parameter and the second model parameter are adapted to a prediction model of the terminal; receiving, by the server, a first prediction loss sent by at least one of the plurality of terminals, where a first prediction loss sent by each of the at least one terminal is calculated by the terminal based on the prediction model that uses the first model parameter and the second model parameter; and updating, by the server, the first model parameter based on the first prediction loss sent by the at least one terminal, to obtain an updated first model parameter, and updating the second model parameter based on the first prediction loss sent by the at least one terminal, to obtain an updated second model parameter. According to the embodiments of the present invention, a prediction model that reflects a user difference can be obtained through training by using a large amount of user data.

Machine learning model training method and device

Embodiments of the present invention provide a machine learning model training method. The method includes: obtaining target task training data and N categories of support task training data; inputting the target task training data and the N categories of support task training data into a memory model to obtain target task training feature data and N categories of support task training feature data; training the target task model based on the target task training feature data and obtaining a first loss of the target task model, and separately training respectively corresponding support task models based on the N categories of support task training feature data and obtaining respective second losses of the N support task models; and updating the memory model, the target task model, and the N support task models based on the first loss and the respective second losses of the N support task models. Training a machine learning model according to the method provided in the embodiments of the present invention can improve accuracy of a prediction result.

Meta-learning for camera adaptive color constancy

A processing entity configured to generate a model for estimating scene illumination colour for a source image captured by a camera, by the steps of: acquiring a set of images, each image having been captured by a respective camera, the set of images as a whole comprising images captured by multiple cameras; forming a set of tasks by assigning each image of the set of images to a respective task such that images in the same task have in common that a property of those images lies in a predetermined range; training parameters of the model by repeatedly: selecting at least one of the tasks, forming an interim set of model parameters in dependence on a first subset of the images of that task, estimating the quality of the interim set of model parameters against a second subset of the images of that task and updating the parameters of the model in dependence on the interim set of parameters and the estimated quality.

Method and Apparatus for Determining Neural Network

This application provides a method and related apparatus for determining a neural network in the field of artificial intelligence. The method includes: obtaining a plurality of initial search spaces; determining M candidate neural networks based on the plurality of initial search spaces, where the candidate neural network includes a plurality of candidate subnetworks, the plurality of candidate subnetworks belong to the plurality of initial search spaces, and any two of the plurality of candidate subnetworks belong to different initial search spaces; evaluating the M candidate neural networks to obtain M evaluation results; and determining N candidate neural networks from the M candidate neural networks based on the M evaluation results, and determining N first target neural networks based on the N candidate neural networks. According to the method and the related apparatus provided in this application, a combined neural network with relatively high performance can be obtained.

Machine learning model training method and device

Oligosaccharide compound for inhibiting endogenous coagulation factor x-enzyme complex, and preparation method therefor and uses thereof

A purified oligosaccharide compound having antithrombotic activity or a mixture of a homologous compound thereof and a pharmaceutically acceptable salt thereof, a preparation method for the mixture, a pharmaceutical composition containing the mixture, and uses thereof serving as an intrinsic factor X-enzyme (Xase) inhibitor in the preparation of drugs for preventing and/or treating thrombotic diseases.

Recommendation method and apparatus based on automatic feature grouping

This application relates to the field of artificial intelligence. A recommendation method based on automatic feature grouping includes: obtaining a plurality of candidate recommended objects and a plurality of association features of each of the plurality of candidate recommended objects; performing multi-order automatic feature grouping on the plurality of association features of each candidate recommended object, to obtain a multi-order feature interaction set of each candidate recommended object; obtaining an interaction feature contribution value of each candidate recommended object through calculation based on the plurality of association features in the multi-order feature interaction set of each candidate recommended object; obtaining a prediction score of each candidate recommended object through calculation based on the interaction feature contribution value of each candidate recommended object; and determining one or more corresponding candidate recommended objects with a high prediction score as a target recommended object.

Distributed graph computation method and device

Embodiments of the present invention disclose a graph partitioning method. The method in the embodiments of the present invention includes: after a graph partitioning apparatus extracts an edge, first determining whether an aggregation degree between a currently extracted edge and an allocated edge in a first device satisfies a preset condition; then, when the preset condition is satisfied, determining whether a quantity of allocated edges stored in the first device is less than a first preset threshold; and allocating the currently extracted edge to the first device when the quantity is less than the first preset threshold. In this way, an aggregation degree between allocated edges in each device is relatively high and each device has relatively balanced load. When an edge changes and an edge associated with the particular edge needs to be synchronized, a relatively small quantity of devices need to perform synchronization and update, so that costs of communication between devices are reduced, and distributed graph computing efficiency is improved.

Deep learning training method and apparatus for use in computing device

Meta-learning for camera adaptive color constancy

Neural network construction method and apparatus

Image classification method and related device thereof

This application provides an image classification method and a related device thereof, applied to the artificial intelligence field, to accurately identify an image belonging to a new category, and not to determine the image belonging to the new category as an image belonging to a basic category by mistake, thereby improving accuracy of image classification. The method of this application includes: obtaining a first feature of a reference image and a second feature of a to-be-classified image; generating a third feature based on the first feature and the second feature; generating a first classification result based on the first feature, where the first classification result is used to determine a category of the reference image; generating a second classification result based on the third feature; and generating a third classification result based on the first classification result and the second classification result, where the third classification result is used to determine a category of the to-be-classified image.

Neural network training method, image classification system, and related device

Method, apparatus and system for generating neural network model, devices, medium and program product

Embodiments of the present disclosure provide a method, an apparatus, and a system for generating a neural network model, a device, a medium, and a program product. The method includes: A first device sends an indication about a structure of a subnetwork model to a second device, where the subnetwork model is determined by adjusting a structure of a hypemetwork model. The first device receives a parameter of the subnetwork model from the second device, where the parameter of the subnetwork model is determined by the second device based on the indication and the hypemetwork model. The first device trains the subnetwork model based on the received parameter of the subnetwork model. The first device sends a parameter of the trained subnetwork model to the second device for the second device to update the hypemetwork model. In the foregoing manner, an efficient federated learning scheme between a plurality of devices is provided, so that communication costs and device computing costs required in a federated learning process are further reduced while improving model precision.

Data processing method and device

This application provides a data processing method and an apparatus. This application relates to the field of artificial intelligence, and specifically, to the field of computer vision. The method includes: adding an architecture parameter to each feature interaction item in a first model, to obtain a second model, where the first model is an FM-based model, and the architecture parameter represents importance of a corresponding feature interaction item; performing optimization on architecture parameters in the second model, to obtain the optimized architecture parameters; obtaining, based on the optimized architecture parameters and the first model or the second model, a third model through feature interaction item deletion. Feature interaction item selection can be performed through optimization on the architecture parameters, and training for a plurality of candidate subsets in a conventional technology is not required. Therefore, this can effectively reduce a computing workload of feature interaction selection to save computing power, and improve efficiency of feature interaction selection.

Molecular sieves with intergrown phases of aei and cha topologies and catalyst thereof

The present disclosure provides molecular sieves with intergrown phases of AEI and CHA topologies and a catalyst thereof. A preparation method for the molecular sieves include the following steps: mixing a hydroxyphosphono organic alkali R with an aluminum source and a silicon source to obtain a sol-gel precursor, putting the sol-gel precursor into a closed hydrothermal synthesis reactor for reaction, filtering the reaction solution, washing, drying, and calcination to obtain the molecular sieves with intergrown phases of AEI and CHA topologies. The molecular sieves and the catalyst thereof can be directly synthesized under mild conditions with a hydroxyphosphono organic alkali as a structure-directing agent and a phosphorus source, have a pH value of 6-9 and low requirements for corrosion resistance of production devices, and are suitable for large-scale production.

Oligosaccharide Compound for Inhibiting Intrinsic Coagulation Factor X-Enzyme Complex, and Preparation Method Therefor and Uses Thereof

A purified oligosaccharide compound having antithrombotic activity or a mixture of a homologous compound thereof and a pharmaceutically acceptable salt thereof, a preparation method for the mixture, a pharmaceutical composition containing the mixture, and uses thereof serving as an intrinsic factor X-enzyme (Xase) inhibitor in the preparation of drugs for preventing and/or treating thrombotic diseases.

Data processing method and apparatus

A data processing method, includes receiving a data flow; generating a triplet set according to the data flow, where each triplet in the set includes three items, the first item is a first element in the data flow, the second item includes a first time point at which the first element appears in the data flow and a first quantity of times that corresponds to the first time point, and the third item includes a second time point at which the first element appears in the data flow and a second quantity of times that corresponds to the second time point; and performing data processing on the data flow according to the triplet set. In the embodiments of the present application, the triplet set may be generated based on the data flow.

ＯＦＦ－ＥＲＩ構造のｍｓｅｃｔ－４モレキュラーシーブ、その製造方法及び使用

【課題】水熱安定性に優れ、規則的な棒状構造の露出する吸着サイトが多く、比表面積が大きく、選択的接触還元、受動的吸着及び接触分解の複数の技術分野に適用可能なモレキュラーシーブ、その製造方法及び使用を提供する。【解決手段】８員環小孔モレキュラーシーブを原料とし、水相に分散させ、苛性カリ、アルミニウム源及び構造規定剤を加えてｐＨを１０超に調整する。ケイ素源を加えて、Ｓｉ／Ａｌ比を調整し、撹拌反応、熟成及び結晶化を経た後、濾過、洗浄、アンモウム交換反応、乾燥及び焼成を行う。【選択図】図１

Normalization in deep convolutional neural networks

A device for machine learning is provided, including a first neural network layer, a second neural network layer with a normalization layer arranged in between. The normalization layer is configured to, when the device is undergoing training on a batch of training samples, receive multiple outputs of the first neural network layer for a plurality of training samples of the batch, each output comprising multiple data values for different indices on a first dimension and a second dimension; group the outputs into multiple groups based on the indices on the first and second dimensions; form a normalization output for each group which are provided as input to the second neural network layer. According to the application, the training of a deep convolutional neural network with good performance that performs stably at different batch sizes and is generalizable to multiple vision tasks is achieved, thereby improving the performance of the training.

Data coding method and related device

This application relates to the field of artificial intelligence, and discloses a data encoding method. The method includes: obtaining to-be-encoded data; processing the to-be-encoded data by using a volume preserving flow model, to obtain a hidden variable output, where the volume preserving flow model includes a target volume preserving flow layer, an operation corresponding to the target volume preserving flow layer is an invertible operation that meets a volume preserving flow constraint, the target volume preserving flow layer is used to perform a multiplication operation on a preset coefficient and first data input to the target volume preserving flow layer, and the preset coefficient is not 1; and encoding the hidden variable output, to obtain encoded data. In this application, lossless compression is implemented by using the volume preserving flow model. On a premise that the target volume preserving flow layer of the volume preserving flow model is invertible, the operation corresponding to the target volume preserving flow layer includes an operation (multiplication operation) in addition to integer addition and subtraction operations, so that the volume preserving flow model has a stronger representation capability. This can more accurately determine data distribution, and implement a better compression rate.

Data Encoding Method and Related Device

A data encoding method includes obtaining to-be-encoded data; processing the to-be-encoded data by using a volume preserving flow model to obtain a hidden variable output, where the volume preserving flow model includes a target volume preserving flow layer, an operation corresponding to the target volume preserving flow layer is an invertible operation that meets a volume preserving flow constraint, the target volume preserving flow layer is used to perform a multiplication operation on a preset coefficient and first data input to the target volume preserving flow layer, and the preset coefficient is not 1; and encoding the hidden variable output to obtain encoded data.

Neural network searching method and related device

This application relates to the artificial intelligence field, and discloses a neural network search method and a related apparatus. The neural network search method includes: constructing attention heads (heads) in transformer layers by sampling a plurality of candidate operators during model search, to construct a plurality of candidate neural networks, and comparing performance of the plurality of candidate neural networks to select a target neural network with higher performance. In this application, a transformer model is constructed with reference to model search, so that a new attention structure with better performance than an original self-attention mechanism can be generated, and effect in a wide range of downstream tasks is significantly improved.

Synthesis method for directly preparing h-type molecular sieve having cha framework type and use thereof

An H-type molecular sieve having the CHA framework type is directly prepared through hydrothermal synthesis by using the compounds with the chemical formulas (C4H8NO)3PO and/or (C4H8NO)2PO(C4H10N) as structure-directing agents. The structure-directing agent has a molecular topological structure, which is conducive to quickly and efficiently building a molecular sieve having the CHA framework type, so that the prepared product has a regular morphology and high relative crystallinity. The synthesis method can directly synthesize the H-type molecular sieve having the CHA framework type with a relative crystallinity greater than 95%. The molecular sieve product can be obtained by direct drying and calcination without an ammonium exchange process, so the synthesis method is simple, easy to perform, and suitable for large-scale production. The samarium-based CHA molecular sieve catalyst shows excellent ammonia adsorption capacity, low-temperature activity, active temperature window, N2 selectivity, and structural stability in the NH3-SCR technology.

Neural network search method and related device

This application relates to the artificial intelligence field, and discloses a neural network search method and a related apparatus. The neural network search method includes: constructing attention heads in transformer layers by sampling a plurality of candidate operators during model search, to construct a plurality of candidate neural networks, and comparing performance of the plurality of candidate neural networks to select a target neural network with higher performance. In this application, a transformer model is constructed with reference to model search, so that a new attention structure with better performance than an original self-attention mechanism can be generated, and effect in a wide range of downstream tasks is significantly improved.

Neural network building method and apparatus

A neural network building method and apparatus are disclosed, and relate to the field of artificial intelligence. The method includes: initializing a search space and a plurality of building blocks, where the search space includes a plurality of operators, and the building block is a network structure obtained by connecting a plurality of nodes by using the operator; during training, in at least one training round, randomly discarding some operators, and updating the plurality of building blocks by using operators that are not discarded; and building a target neural network based on the plurality of updated building blocks. In the method, some operators are randomly discarded. This breaks association between operators, and overcomes a co-adaptation problem during training, to obtain a target neural network with better performance.

Data processing method and apparatus

Method, apparatus, and system for generating neural network model, device, medium, and program product

A method, an apparatus, and a system for generating a neural network model, a device, a medium, and a program product are provided. In an embodiment, a first device sends an indication about a structure of a subnetwork model to a second device, where the subnetwork model is determined by adjusting a structure of a hypernetwork model. The first device receives a parameter of the subnetwork model from the second device, where the parameter of the subnetwork model is determined by the second device based on the indication and the hypernetwork model. The first device trains the subnetwork model based on the received parameter of the subnetwork model. The first device sends a parameter of the trained subnetwork model to the second device for the second device to update the hypernetwork model. In the foregoing manner, an efficient federated learning scheme between a plurality of devices is provided.

Automatic driving method, training method and related apparatuses

Recommendation method and device based on automatic feature grouping

This application relates to the field of artificial intelligence. A recommendation method based on automatic feature grouping includes: obtaining a plurality of candidate recommended objects and a plurality of association features of each of the plurality of candidate recommended objects; performing multi-order automatic feature grouping on the plurality of association features of each candidate recommended object, to obtain a multi-order feature interaction set of each candidate recommended object; obtaining an interaction feature contribution value of each candidate recommended object through calculation based on the plurality of association features in the multi-order feature interaction set of each candidate recommended object; obtaining a prediction score of each candidate recommended object through calculation based on the interaction feature contribution value of each candidate recommended object; and determining one or more corresponding candidate recommended objects with a high prediction score as a target recommended object. The recommendation method is applicable to various recommendation-related application scenarios, and helps improve recommendation efficiency and accuracy.

Recommendation method and device based on automatic feature grouping

Similarity measurement method and equipment

Lane line detection method, and related apparatus

Method and device for constructing neural network

Molecular sieves with intergrown phases of AEI and CHA topologies and catalyst thereof

The present disclosure provides molecular sieves with intergrown phases of AEI and CHA topologies and a catalyst thereof. A preparation method for the molecular sieves include the following steps: mixing a hydroxyphosphono organic alkali R with an aluminum source and a silicon source to obtain a sol-gel precursor, putting the sol-gel precursor into a closed hydrothermal synthesis reactor for reaction, filtering the reaction solution, washing, drying, and calcination to obtain the molecular sieves with intergrown phases of AEI and CHA topologies. The molecular sieves and the catalyst thereof can be directly synthesized under mild conditions with a hydroxyphosphono organic alkali as a structure-directing agent and a phosphorus source, have a pH value of 6-9 and low requirements for corrosion resistance of production devices, and are suitable for large-scale production.

Msect-4 molecular sieves with off and eri topologies, preparation method therefor, and applications thereof

The present disclosure provides msect-4 molecular sieves with OFF and ERI topologies, a preparation method therefor, and applications thereof. An eight-membered ring small pore molecular sieve used as a raw material is dispersed in an aqueous phase. Following that, caustic potash, an aluminum source, and an organic structure-directing agent (OSDA) are added. The pH value is then adjusted to be greater than 10, and a silicon source is introduced to attain the desired silicon-aluminum ratio, followed by stirring reaction, aging, crystallization, filtration, washing, ammonia exchange reaction, drying, and calcination. The msect-4 molecular sieves with OFF and ERI topologies, the preparation method therefor, and applications exhibit excellent hydrothermal stability, a plurality of adsorption sites exposed by a regular bone-like structure, and a large specific surface area. Consequently, this molecular sieves find applicability across various technical fields including selective catalytic reduction, passive adsorption, and catalytic cracking, and has broad application prospects.

MSECT-4 molecular sieves with off and ERI topologies, preparation method therefor, and applications thereof

The present disclosure provides msect-4 molecular sieves with OFF and ERI topologies, a preparation method therefor, and applications thereof. An eight-membered ring small pore molecular sieve used as a raw material is dispersed in an aqueous phase. Following that, caustic potash, an aluminum source, and an organic structure-directing agent (OSDA) are added. The pH value is then adjusted to be greater than 10, and a silicon source is introduced to attain the desired silicon-aluminum ratio, followed by stirring reaction, aging, crystallization, filtration, washing, ammonia exchange reaction, drying, and calcination. The msect-4 molecular sieves with OFF and ERI topologies, the preparation method therefor, and applications exhibit excellent hydrothermal stability, a plurality of adsorption sites exposed by a regular bone-like structure, and a large specific surface area. Consequently, this molecular sieves find applicability across various technical fields including selective catalytic reduction, passive adsorption, and catalytic cracking, and has broad application prospects.

System and method for cross-modal interaction based on pre-trained model

A method is provided for data processing performed by a processing system. The method comprises determining a set of first tokens for first data and a set of second token for second data, each token comprising information associated with a segment of the respective data, determining pair-wise similarities between the set of first tokens and the set of second tokens, each pair comprising a first token in the set of first tokens and a second token in the set of second tokens, determining, for each first token in the set of first tokens, a maximum similarity based on the determined pair-wise similarities between the respective first token and the second tokens in the set of second tokens, and determining a first similarity between the first data and the second data by aggregating the maximum similarities corresponding to the first tokens in the set of first set of tokens.

Mutant p-hydroxyphenylpyruvate dioxygenase, and coding nucleic acid and use thereof

Disclosed are a mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein, a biologically active fragment thereof and an isolated polynucleotide comprising a nucleic acid sequence encoding the protein or fragment, wherein the mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein or the biologically active fragment thereof retains or enhances its property of catalyzing the conversion of p-hydroxyphenylpyruvic acid (HPP) into homogentisic acid, and is less sensitive to HPPD inhibitors than wild-type HPPD. Also disclosed are a nucleic acid construct, expression vector and host cell comprising the polynucleotide, as well as a method for producing a plant that has the property of catalyzing the conversion of p-hydroxyphenylpyruvic acid (HPP) into homogentisic acid, and at the same time, has a reduced sensitivity to HPPD inhibitory herbicides.

Image Classification Method and Related Device Thereof

An image classification method includes obtaining a first feature of a reference image and a second feature of a to-be-classified image; generating a third feature based on the first feature and the second feature; generating a first classification result based on the first feature, where the first classification result is used to determine a category of the reference image; generating a second classification result based on the third feature; and generating a third classification result based on the first classification result and the second classification result, where the third classification result is used to determine a category of the to-be-classified image.

Data processing method and device

Data processing method and apparatus

Embodiments of this application provide a data processing method and apparatus to better learn a vector representation value of each feature value in a continuous feature. The method specifically includes: The data processing apparatus obtains the continuous feature from sample data, and then performs discretization processing on the continuous feature by using a discretization model, to obtain N discretization probabilities corresponding to the continuous feature. The N discretization probabilities correspond to N preset meta-embeddings, and N is an integer greater than 1. Finally, the data processing apparatus determines a vector representation value of the continuous feature based on the N discretization probabilities and the N meta-embeddings.

System and method for cross-modal interaction based on pre-trained model

A method is provided for data processing performed by a processing system. The method comprises determining a set of first tokens for first data and a set of second token for second data, each token comprising information associated with a segment of the respective data, determining pair-wise similarities between the set of first tokens and the set of second tokens, each pair comprising a first token in the set of first tokens and a second token in the set of second tokens, determining, for each first token in the set of first tokens, a maximum similarity based on the determined pair-wise similarities between the respective first token and the second tokens in the set of second tokens, and determining a first similarity between the first data and the second data by aggregating the maximum similarities corresponding to the first tokens in the set of first set of tokens.

Data encoding method, data decoding method, and data processing apparatus

This application relates to the field of artificial intelligence, and discloses a data encoding method, a data decoding method, and data processing apparatuses. Both the data encoding method and the data decoding method relate to an invertible flow-based model. The invertible flow-based model includes a target invertible flow layer, a model parameter of the target invertible flow layer is used to constrain an auxiliary variable generated in an inverse transform processing process, an operation corresponding to the target invertible flow layer includes a multiplication operation and a division operation that are determined based on the model parameter, and the auxiliary variable is an increment of a product of the multiplication operation or a remainder generated through the division operation. In processes of data encoding and data decoding, not only the multiplication operation can be used, but invertibility can be implemented by using the auxiliary variable when the division operation is performed. Therefore, in comparison with an integer discrete flow model in the conventional technology, a model representation capability is improved. This helps improve estimation accuracy of model probability distribution, and can further improve compression efficiency of lossless compression.

P-hidroxifenilpiruvato dioxigenasa mutante, ácido nucleico que la codifica y su uso.

La presente invención se refiere a una proteína p–hidroxifenilpiruvato dioxigenasa (HPPD) mutante o un fragmento bioactivo de la misma y un polinucleótido aislado que comprende una secuencia de ácido nucleico que codifica la proteína o fragmento de la misma, en donde la proteína p– hidroxifenilpiruvato dioxigenasa (HPPD) o un fragmento bioactivo de la misma conserva o mejora la propiedad de catalizar la conversión de p– hidroxifenilpiruvato (HPP) en homogentizado y es significativamente menos sensible a los herbicidas inhibidores de HPPD que una HPPD de tipo salvaje. La presente invención también se refiere a un constructo de ácido nucleico, un vector de expresión y una célula huésped que comprende el polinucleótido, así como a un método para producir una planta que tiene la propiedad de catalizar la conversión de p–hidroxifenilpiruvato (HPP) a homogentizado y sensibilidad significativamente reducida a los <br /> herbicidas inhibidores de HPPD.

Mutant p-hydroxyphenylpyruvate dioxygenase, nucleic acid encoding the same and use thereof

Disclosed are a mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein, a biologically active fragment thereof and an isolated polynucleotide comprising a nucleic acid sequence encoding the protein or fragment, wherein the mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein or the biologically active fragment thereof retains or enhances its property of catalyzing the conversion of p-hydroxyphenylpyruvic acid (HPP) into homogentisic acid, and is less sensitive to HPPD inhibitors than wild-type HPPD. Also disclosed are a nucleic acid construct, expression vector and host cell comprising the polynucleotide, as well as a method for producing a plant that has the property of catalyzing the conversion of p-hydroxyphenylpyruvic acid (HPP) into homogentisic acid, and at the same time, has a reduced sensitivity to HPPD inhibitory herbicides.

p-hidroxifenilpiruvato dioxigenase mutante, ácido nucleico codificando a mesma, e uso dos mesmos

A presente invenção refere-se a uma proteína p-hidroxifenilpiruvato dioxigenase (HPPD) mutante ou um fragmento bioativo da mesma e um polinucleotídeo isolado compreendendo uma sequência de ácido nucleico que codifica a proteína ou fragmento da mesma, em que a proteína p-hidroxifenilpiruvato dioxigenase (HPPD) mutante ou um fragmento bioativo da mesma mantém ou aumenta a propriedade de catalisar a conversão de p-hidroxifenilpiruvato (HPP) em homogentisato e é significativamente menos sensível a herbicidas inibidores de HPPD do que HPPD de ocorrência natural. A presente invenção também se refere a um construto de ácido nucleico, um vetor de expressão e uma célula hospedeira compreendendo o polinucleotídeo, bem como a um método para produzir uma planta que tem a propriedade de catalisar a conversão de p-hidroxifenilpiruvato (HPP) em homogentisato e sensibilidade significativamente reduzida aos herbicidas inibidores de HPPD.

Neural network obtaining method, data processing method, and related device

A neural network obtaining method, a data processing method, and a related device are disclosed. The disclosed methods may be used in the field of automatic neural architecture search technologies in the field of artificial intelligence. An example method includes: obtaining first indication information, where the first indication information indicates a probability and/or a quantity of times that k neural network modules appear in a first neural architecture cell; generating the first neural architecture cell based on the first indication information, and generating a first neural network; obtaining a target score corresponding to the first indication information, where the target score indicates performance of the first neural network; and obtaining second indication information from a plurality of pieces of first indication information based on a plurality of target scores, and obtaining a target neural network corresponding to the second indication information.

Data processing method and apparatus

Embodiments of this application provide a data processing method and apparatus to better learn a vector representation value of each feature value in a continuous feature. The method specifically includes: The data processing apparatus obtains the continuous feature from sample data, and then performs discretization processing on the continuous feature by using a discretization model, to obtain N discretization probabilities corresponding to the continuous feature. The N discretization probabilities correspond to N preset meta-embeddings, and N is an integer greater than 1. Finally, the data processing apparatus determines a vector representation value of the continuous feature based on the N discretization probabilities and the N meta-embeddings.

Image classification method and apparatus, and image classification model training method and apparatus

An image classification method and apparatus, and an image classification model training method and apparatus are provided, which are related to an image recognition technology in the artificial intelligence field and more specifically, to the computer vision field. The method includes: obtaining a to-be-processed image; and classifying the to-be-processed image based on a preset global class feature, to obtain a classification result of the to-be-processed image. The preset global class feature includes a plurality of class features obtained through training based on a plurality of training images in a training set. The plurality of class features in the preset global class feature are used to indicate visual features of all classes in the training set.

Neural network acquisition method, data processing method and related device

A neural network obtaining method, a data processing method, and a related device are disclosed. The method may be used in the field of automatic neural architecture search technologies in the field of artificial intelligence. The method includes: obtaining first indication information, where the first indication information indicates a probability and/or a quantity of times that k neural network modules appear in a first neural architecture cell; generating the first neural architecture cell based on the first indication information, and generating a first neural network; obtaining a target score corresponding to the first indication information, where the target score indicates performance of the first neural network; and obtaining second indication information from a plurality of pieces of first indication information based on a plurality of target scores, and obtaining a target neural network corresponding to the second indication information. The first indication information no longer indicates a topology relationship between different neural network modules. Therefore, search space corresponding to the neural architecture cell is reduced, required computer resources are reduced, and time costs are reduced.

Data coding method and related device

Data encoding method, data decoding method, and data processing apparatus

This application relates to the field of artificial intelligence, and discloses a data encoding method, a data decoding method, and data processing apparatuses. Both the data encoding method and the data decoding method relate to an invertible flow-based model. The invertible flow-based model includes a target invertible flow layer, a model parameter of the target invertible flow layer is used to constrain an auxiliary variable generated in an inverse transform processing process, an operation corresponding to the target invertible flow layer includes a multiplication operation and a division operation that are determined based on the model parameter, and the auxiliary variable is an increment of a product of the multiplication operation or a remainder generated through the division operation.

Processing apparatus, method and storage medium

Mutant p-hydroxyphenylpyruvate dioxygenase, and coding nucleic acid and use thereof

Disclosed are a mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein, a biologically active fragment thereof and an isolated polynucleotide comprising a nucleic acid sequence encoding the protein or fragment, wherein the mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein or the biologically active fragment thereof retains or enhances its property of catalyzing the conversion of p-hydroxyphenylpyruvic acid (HPP) into homogentisic acid, and is less sensitive to HPPD inhibitors than wild-type HPPD. Also disclosed are a nucleic acid construct, expression vector and host cell comprising the polynucleotide, as well as a method for producing a plant that has the property of catalyzing the conversion of p-hydroxyphenylpyruvic acid (HPP) into homogentisic acid, and at the same time, has a reduced sensitivity to HPPD inhibitory herbicides.

Mutant p-hydroxyphenylpyruvate dioxygenase, and coding nucleic acid and use thereof

The present invention relates to a mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein or a bioactive fragment thereof and an isolated polynucleotide comprising a nucleic acid sequence encoding the protein or fragment thereof, wherein the mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein or a bioactive fragment thereof retains or enhances the property of catalyzing the conversion of p-hydroxyphenylpyruvate (HPP) to homogentisate and is significantly less sensitive to HPPD-inhibiting herbicides than a wild-type HPPD. The present invention also relates to a nucleic acid construct, an expression vector and a host cell comprising the polynucleotide, as well as to a method for producing a plant that has the property of catalyzing the conversion of p-hydroxyphenylpyruvate (HPP) to homogentisate and significantly reduced sensitivity to HPPD-inhibiting herbicides.

Normalization in deep convolutional neural networks

Neural network construction method and system

Neural network optimization method and device

Method and apparatus for searching for neural network ensemble model, and electronic device

Disclosed is a method for searching for a neural network architecture ensemble model. The method includes: obtaining a dataset, where the dataset includes a sample and an annotation in a classification task; performing search by using a distributional neural network architecture search algorithm, including: determining a hyperparameter of a neural network architecture distribution; sampling a valid (valid) neural network architecture from the architecture distribution defined by the hyperparameter; training and evaluating the neural network architecture on the dataset, to obtain a performance indicator; determining, based on the performance indicator, neural network architecture distributions that share the hyperparameter, to obtain a candidate pool of base learners, where the base learner is a neural network architecture that meets an architecture distribution requirement; and determining a surrogate model, where the surrogate model is used to predict test performance of an unevaluated neural network architecture; and predicting test performance of the base learner in the candidate pool by using the surrogate model, and determining that k diverse (diverse) base learners that meet a task scenario requirement form an ensemble model, where a size of the ensemble model is k .

Mutant p-hydroxyphenylpyruvate dioxygenase, and coding nucleic acid and use thereof

Disclosed are a mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein, a biologically active fragment thereof and an isolated polynucleotide comprising a nucleic acid sequence encoding the protein or fragment, wherein the mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein or the biologically active fragment thereof retains or enhances its property of catalyzing the conversion of p-hydroxyphenylpyruvic acid (HPP) into homogentisic acid, and is less sensitive to HPPD inhibitors than wild-type HPPD. Also disclosed are a nucleic acid construct, expression vector and host cell comprising the polynucleotide, as well as a method for producing a plant that has the property of catalyzing the conversion of p-hydroxyphenylpyruvic acid (HPP) into homogentisic acid, and at the same time, has a reduced sensitivity to HPPD inhibitory herbicides.

Mutant p-hydroxyphenylpyruvate dioxygenase, and coding nucleic acid and use thereof

The present invention relates to a mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein or a bioactive fragment thereof and an isolated polynucleotide comprising a nucleic acid sequence encoding the protein or fragment thereof, wherein the mutant p-hydroxyphenylpyruvate dioxygenase (HPPD) protein or a bioactive fragment thereof retains or enhances the property of catalyzing the conversion of p-hydroxyphenylpyruvate (HPP) to homogentisate and is significantly less sensitive to HPPD-inhibiting herbicides than a wild-type HPPD. The present invention also relates to a nucleic acid construct, an expression vector and a host cell comprising the polynucleotide, as well as to a method for producing a plant that has the property of catalyzing the conversion of p-hydroxyphenylpyruvate (HPP) to homogentisate and significantly reduced sensitivity to HPPD-inhibiting herbicides.

Method for data processing by performing different non-linear combination processing

The method includes: obtaining a plurality of pieces of feature data; automatically performing two different types of nonlinear combination processing operations on the plurality of pieces of feature data to obtain two groups of processed data, where the two groups of processed data include a group of higher-order data and a group of lower-order data, the higher-order data is related to a nonlinear combination of m pieces of feature data in the plurality of pieces of feature data, and the lower-order data is related to a nonlinear combination of n pieces of feature data in the plurality of pieces of feature data, where m≥3, and m>n≥2; and determining prediction data based on a plurality of pieces of target data, where the plurality of pieces of target data include the two groups of processed data.

Method and apparatus for searching for neural network ensemble model, and electronic device

Disclosed is a method for searching for a neural network architecture ensemble model. The method includes: obtaining a dataset, where the dataset includes a sample and an annotation in a classification task; performing search by using a distributional neural network architecture search algorithm, including: determining a hyperparameter of a neural network architecture distribution; sampling a valid neural network architecture from the architecture distribution defined by the hyperparameter; training and evaluating the neural network architecture on the dataset, to obtain a performance indicator; determining, based on the performance indicator, neural network architecture distributions that share the hyperparameter, to obtain a candidate pool of base learners; and determining a surrogate model; and predicting test performance of the base learner in the candidate pool by using the surrogate model, and determining that k diverse base learners that meet a task scenario requirement form an ensemble model.

Image classification method and related device thereof

Supported platinum-based three-way catalyst regeneration method

A method for improving the sintering resistance of noble metal nanocatalysts

Machine learning model training method and apparatus

Embodiments of the present invention provide a machine learning model training method, including: obtaining target task training data and N categories of support task training data; inputting the target task training data and the N categories of support task training data into a memory model to obtain target task training feature data and N categories of support task training feature data; training the target task model based on the target task training feature data and obtaining a first loss of the target task model, and separately training respectively corresponding support task models based on the N categories of support task training feature data and obtaining respective second losses of the N support task models; and updating the memory model, the target task model, and the N support task models based on the first loss and the respective second losses of the N support task models.

Neural network searching method and related device

Method, apparatus and system for generating neural network model, devices, medium and program product