CN111767981A - Approximate calculation method of Mish activation function - Google Patents
Approximate calculation method of Mish activation function Download PDFInfo
- Publication number
- CN111767981A CN111767981A CN202010431133.XA CN202010431133A CN111767981A CN 111767981 A CN111767981 A CN 111767981A CN 202010431133 A CN202010431133 A CN 202010431133A CN 111767981 A CN111767981 A CN 111767981A
- Authority
- CN
- China
- Prior art keywords
- function
- activation function
- neural network
- mish
- convolutional neural
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000004913 activation Effects 0.000 title claims abstract description 92
- 238000004364 calculation method Methods 0.000 title claims abstract description 62
- 238000000034 method Methods 0.000 claims abstract description 14
- 230000008569 process Effects 0.000 claims abstract description 8
- 238000013527 convolutional neural network Methods 0.000 claims description 45
- 101150060512 SPATA6 gene Proteins 0.000 claims description 15
- 238000013528 artificial neural network Methods 0.000 claims description 8
- 230000011218 segmentation Effects 0.000 claims description 8
- 230000000007 visual effect Effects 0.000 claims description 6
- 238000009499 grossing Methods 0.000 claims description 3
- 238000012886 linear function Methods 0.000 claims description 3
- 238000004088 simulation Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 abstract description 148
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Complex Calculations (AREA)
- Image Analysis (AREA)
Abstract
The invention provides an approximate calculation method of a Mish activation function, which constructs and forms a hard-Mish piecewise function with a simpler form by adopting a piecewise approximation mode, wherein the calculation complexity of the hard-Mish piecewise function is far lower than that of the Mish activation function, so that the time consumption of function operation can be effectively reduced, in addition, the hard-Mish piecewise function can also effectively reduce the access times and the memory occupancy rate of a system memory in the operation process, and the calculation results of the hard-Mish piecewise function and the Mish activation function have smaller errors, so that the application universality of the Mish activation function is effectively improved.
Description
Technical Field
The invention relates to the technical field of computer data processing, in particular to an approximate calculation method of a Mish activation function.
Background
The Mish activation function is a newly researched activation function, and the specific mathematical form of the Mish activation function is Mish (x) -x-tanh (ln (1+ e)x) Compared with the prior sigmoid and swish activation functions, the Mish activation function has more excellent calculation performance in the aspects of detection and identification of a plurality of computer vision tasks, but the Mish activation function occupies more system memory and consumes longer calculation time in the operation process due to the complex calculation formula of the Mish activation function, so that the Mish activation function is seriously restrictedThe application universality of the live function. Therefore, the prior art continues to provide a function approximation calculation method which can approximate to the Mish activation function and has a small calculation error with the Mish activation function.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides an approximate calculation method of a Mish activation function, which comprises the following steps: step S1, constructing a convolution neural network; step S2, constructing another hard-hash function according to the piecewise approximation principle about the Mish activation function, and taking the function as a new activation function; step S3, training the convolutional neural network according to the hard-hash function, so as to update the weight parameter of the convolutional neural network to a convergence state; step S4, outputting the convolution neural network with the weight parameter in the convergence state; therefore, the approximate calculation method of the Mish activation function adopts a piecewise approximation mode to construct and form a relatively simple hard-hash piecewise function, the calculation complexity of the hard-hash piecewise function is far lower than that of the Mish activation function, so that the time consumed by function operation can be effectively reduced, in addition, the hard-hash piecewise function can also effectively reduce the access times and the memory occupancy rate of a system memory in the operation process, and the hard-hash piecewise function and the Mish activation function have relatively small errors in the calculation result, so that the application universality of the Mish activation function is effectively improved.
The invention provides an approximate calculation method of a Mish activation function, which is characterized by comprising the following steps of:
step S1, constructing a convolution neural network;
step S2, constructing another hard-hash function according to the piecewise approximation principle about the Mish activation function, and taking the function as a new activation function;
step S3, training the convolutional neural network according to the hard-hash function, so as to update the weight parameters of the convolutional neural network to a convergence state;
step S4, outputting the convolution neural network with the weight parameter in a convergence state;
further, in the step S1, the constructing a convolutional neural network specifically includes,
constructing the convolutional neural network model according to a data calculation scene and/or a calculation data type of the Mish activation function;
further, in the step S2, another hard-hash function is constructed according to the piecewise approximation principle about the hash activation function, and the construction specifically includes, as a new activation function,
step S201, carrying out interval segmentation on the Mish activation function so as to obtain function curve parameters of the Mish activation function in a plurality of different intervals;
step S202, performing infinite approximation calculation processing on the function curve parameter of each interval, and constructing a hard-hash function shown in the following formula (1) according to the infinite approximation calculation processing
Further, in the step S201, the segment segmentation is performed on the mesh activation function, so as to obtain function curve parameters of the mesh activation function in a plurality of different segments specifically includes,
step S2011, setting the mathematical formula of the Mish activation function as shown in the following formula (2)
Mish(x)＝x·tanh(ln(1+ex) (2)；
Step S2012, segmenting the variable x in the Mish activation function according to intervals (— infinity, -3], (-3, 3) and [3, + ∞);
step S2013, calculating function calculation result offset of a function curve corresponding to each of the three intervals of the Mish activation function according to the three interval segmentation results of the variable x, and taking the offset as the function curve parameter;
further, in the step S202, performing infinite approximation calculation processing on the function curve parameter of each of the intervals, and thus constructing the hard-hash function shown in the formula (1) specifically includes,
step S2021, calculating the offset of the function calculation result of the function curve corresponding to each interval, and performing simulation of a linear function or a nonlinear function to correspondingly obtain three subfunctions;
step S2022, performing function connection smoothing processing on the three sub-functions at the section connection point x-3 and the section connection point x-3, so as to construct and form a hard-hash function shown in the formula (1);
further, in the step S3, training the convolutional neural network according to the hard-hash function to update the weight parameter of the convolutional neural network to a convergence state specifically includes, in step S301, performing iterative training on the convolutional neural network for a predetermined number of times according to the hard-hash function to obtain the weight parameter of the convolutional neural network;
step S302, judging whether the current weight parameter of the convolutional neural network is in a convergence interval range;
step S303, if the current weight parameter is determined to be in the convergence interval range, stopping the iterative training of the convolutional neural network;
step S304, if the current weight parameter is determined not to be in the convergence interval range, performing single training on the convolutional neural network according to the hard-hash function until the weight parameter obtained after training is in the convergence interval range;
further, in the step S4, the convolutional neural network whose output weight parameter is in a converged state specifically includes,
and outputting the convolutional neural network with the weight parameters in the convergence state to a corresponding visual calculation task so as to execute corresponding visual calculation analysis.
Compared with the prior art, the approximate calculation method of the Mish activation function comprises the following steps: step S1, constructing a convolution neural network; step S2, constructing another hard-hash function according to the piecewise approximation principle about the Mish activation function, and taking the function as a new activation function; step S3, training the convolutional neural network according to the hard-hash function, so as to update the weight parameter of the convolutional neural network to a convergence state; step S4, outputting the convolution neural network with the weight parameter in the convergence state; therefore, the approximate calculation method of the Mish activation function adopts a piecewise approximation mode to construct and form a relatively simple hard-hash piecewise function, the calculation complexity of the hard-hash piecewise function is far lower than that of the Mish activation function, so that the time consumed by function operation can be effectively reduced, in addition, the hard-hash piecewise function can also effectively reduce the access times and the memory occupancy rate of a system memory in the operation process, and the hard-hash piecewise function and the Mish activation function have relatively small errors in the calculation result, so that the application universality of the Mish activation function is effectively improved.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic overall flow chart of an approximate calculation method of a hash activation function provided by the present invention.
Fig. 2 is a detailed flowchart of step S2 in the approximate calculation method of the hash activation function provided by the present invention.
Fig. 3 is a detailed flowchart of step S3 in the approximate calculation method of the hash activation function according to the present invention.
Fig. 4 is an actual error diagram between the hard-hash function and the hash activation function obtained by the approximate calculation method of the hash activation function provided by the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Fig. 1 is a schematic overall flow chart of an approximate calculation method of a hash activation function according to an embodiment of the present invention. The approximate calculation method of the Mish activation function comprises the following steps:
step S1, constructing a convolution neural network;
step S2, constructing another hard-hash function according to the piecewise approximation principle about the Mish activation function, and taking the function as a new activation function;
step S3, training the convolutional neural network according to the hard-hash function, so as to update the weight parameter of the convolutional neural network to a convergence state;
in step S4, the convolutional neural network with the weight parameter in the converged state is output.
Preferably, in this step S1, constructing a convolutional neural network specifically includes,
and constructing the convolutional neural network model according to the data calculation scene and/or the calculation data type of the Mish activation function.
Preferably, in the step S4, the convolutional neural network whose output weight parameters are in a converged state specifically includes,
and outputting the convolutional neural network with the weight parameters in the convergence state to a corresponding visual calculation task so as to execute corresponding visual calculation analysis.
Fig. 2 is a schematic detailed flow chart of step S2 in the approximate calculation method of the Mish activation function according to the embodiment of the present invention. In this step S2, another hard-hash function is constructed according to the piecewise approximation principle with respect to the hash activation function, which specifically includes as a new activation function,
step S201, carrying out interval segmentation on the Mish activation function so as to obtain function curve parameters of the Mish activation function in a plurality of different intervals;
step S202, performing infinite approximation calculation processing on the function curve parameter of each interval, and constructing a hard-hash function shown in the following formula (1) according to the infinite approximation calculation processing
Preferably, in the step S201, the segment segmenting is performed on the nash activation function, so as to obtain function curve parameters of the nash activation function in a plurality of different segments specifically includes,
step S2011, the mathematical formula for setting the Mish activation function is shown in the following formula (2)
Mish(x)＝x·tanh(ln(1+ex) (2)；
Step S2012, segmenting the variable x in the Mish activation function according to the intervals (— infinity, -3], (-3, 3) and [3, + ∞);
step S2013, according to the three interval segmentation results of the variable x, calculating a function calculation result offset of a function curve corresponding to the mesh activation function in each of the three intervals, and taking the function calculation result offset as the function curve parameter.
Preferably, in step S202, performing an infinite approximation calculation process on the function curve parameter of each interval, and thereby constructing the hard-hash function shown in formula (1) specifically includes,
step S2021, calculating the offset of the function calculation result of the function curve corresponding to each interval, and performing simulation of a linear function or a nonlinear function to correspondingly obtain three subfunctions;
in step S2022, the function connection smoothing process is performed on the three subfunctions at the section connection point x-3 and x-3, thereby constructing and forming the hard-hash function shown in the formula (1).
Fig. 3 is a schematic detailed flow chart of step S3 in the approximate calculation method of the Mish activation function according to the embodiment of the present invention. In the step S3, training the convolutional neural network according to the hard-hash function to update the weight parameters of the convolutional neural network to a convergence state specifically includes,
step S301, performing iterative training on the convolutional neural network for a preset number of times according to the hard-hash function so as to obtain a weight parameter of the convolutional neural network;
step S302, judging whether the current weight parameter of the convolutional neural network is in a convergence interval range;
step S303, if the current weight parameter is determined to be in the convergence interval range, stopping the iterative training of the convolutional neural network;
step S304, if it is determined that the current weight parameter is not in the convergence interval range, performing a single training on the convolutional neural network according to the hard-hash function until the weight parameter obtained after the training is in the convergence interval range.
Referring to fig. 4, it is a graph of an actual error between the hard-hash function and the hash activation function obtained by the approximate calculation method for the hash activation function according to the embodiment of the present invention. As can be seen from this fig. 4, the dashed line represents the curve of the Mish activation function, the solid line represents the curve of the hard-Mish function, both function curves having a certain deviation value over the interval (-5, -1), but which can be maintained within (-0.3, +0.3), while both function curves having a substantially negligible deviation value over the interval [ -1, + ∞), in particular over the interval [ +3, + ∞), both function curves being substantially considered to coincide.
From the content of the above embodiment, it can be known that the approximate calculation method of the hash activation function constructs a relatively simple hard-hash piecewise function by means of piecewise approximation, and the computational complexity of the hard-hash piecewise function is much lower than that of the hash activation function, so that the time consumed by function operation can be effectively reduced, in addition, the hard-hash piecewise function can also effectively reduce the access times and memory occupancy rate to the system memory in the operation process, and the hard-hash piecewise function and the hash activation function have relatively small errors in the calculation result, so that the application universality of the hash activation function is effectively improved.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (7)
- An approximate calculation method of a Mish activation function, characterized in that the approximate calculation method of the Mish activation function comprises the following steps:step S1, constructing a convolution neural network;step S2, constructing another hard-hash function according to the piecewise approximation principle about the Mish activation function, and taking the function as a new activation function;step S3, training the convolutional neural network according to the hard-hash function, so as to update the weight parameters of the convolutional neural network to a convergence state;and step S4, outputting the convolutional neural network with the weight parameters in a convergence state.
- 2. The approximate computation method of a Mish activation function as claimed in claim 1, wherein:in step S1, the constructing a convolutional neural network specifically includes,and constructing the convolutional neural network model according to the data calculation scene and/or the calculation data type of the Mish activation function.
- 3. The approximate computation method of a Mish activation function as claimed in claim 1, wherein:in step S2, another hard-hash function is constructed according to the piecewise approximation principle about the hash activation function, which specifically includes as a new activation function,step S201, carrying out interval segmentation on the Mish activation function so as to obtain function curve parameters of the Mish activation function in a plurality of different intervals;step S202, performing infinite approximation calculation processing on the function curve parameter of each interval, and constructing a hard-hash function shown in the following formula (1) according to the infinite approximation calculation processing
- 4. The approximate computation method of a Mish activation function as claimed in claim 3, wherein:in step S201, the step of performing interval segmentation on the mesh activation function to obtain function curve parameters of the mesh activation function in a plurality of different intervals specifically includes,step S2011, setting the mathematical formula of the Mish activation function as shown in the following formula (2)Mish(x)＝x·tanh(ln(1+ex) (2)；Step S2012, segmenting the variable x in the Mish activation function according to intervals (— infinity, -3], (-3, 3) and [3, + ∞);and step S2013, calculating function calculation result offset of a function curve corresponding to each of the three intervals of the Mish activation function according to the three interval segmentation results of the variable x, and taking the function calculation result offset as the function curve parameter.
- 5. The approximate computation method of a Mish activation function as claimed in claim 4, wherein:in the step S202, performing infinite approximation calculation processing on the function curve parameter of each of the intervals, and constructing the hard-hash function shown in formula (1) specifically includes,step S2021, calculating the offset of the function calculation result of the function curve corresponding to each interval, and performing simulation of a linear function or a nonlinear function to correspondingly obtain three subfunctions;in step S2022, the function connection smoothing process is performed on the three sub-functions at the section connection point x-3 and the section connection point x-3, so as to construct and form the hard-hash function shown in the formula (1).
- 6. The approximate computation method of a Mish activation function as claimed in claim 1, wherein:in the step S3, the training of the convolutional neural network according to the hard-hash function to update the weight parameter of the convolutional neural network to a convergence state specifically includes,step S301, performing iterative training on the convolutional neural network for a preset number of times according to the hard-hash function so as to obtain a weight parameter of the convolutional neural network;step S302, judging whether the current weight parameter of the convolutional neural network is in a convergence interval range;step S303, if the current weight parameter is determined to be in the convergence interval range, stopping the iterative training of the convolutional neural network;and step S304, if the current weight parameter is determined not to be in the convergence interval range, performing single training on the convolutional neural network according to the hard-hash function until the weight parameter obtained after training is in the convergence interval range.
- 7. The approximate computation method of a Mish activation function as claimed in claim 1, wherein:in step S4, the convolutional neural network whose output weight parameters are in a converged state specifically includes,and outputting the convolutional neural network with the weight parameters in the convergence state to a corresponding visual calculation task so as to execute corresponding visual calculation analysis.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010431133.XA CN111767981B (en) | 2020-05-20 | 2020-05-20 | Approximate calculation method of Mish activation function |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010431133.XA CN111767981B (en) | 2020-05-20 | 2020-05-20 | Approximate calculation method of Mish activation function |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111767981A true CN111767981A (en) | 2020-10-13 |
| CN111767981B CN111767981B (en) | 2023-12-05 |
Family
ID=72719522
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010431133.XA Active CN111767981B (en) | 2020-05-20 | 2020-05-20 | Approximate calculation method of Mish activation function |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111767981B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022156475A1 (en) * | 2021-01-20 | 2022-07-28 | 华为技术有限公司 | Neural network model training method and apparatus, and data processing method and apparatus |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101404071A (en) * | 2008-11-07 | 2009-04-08 | 湖南大学 | Electronic circuit fault diagnosis neural network method based on grouping particle swarm algorithm |
| CN104408302A (en) * | 2014-11-19 | 2015-03-11 | 北京航空航天大学 | Bearing variable-condition fault diagnosis method based on LMD-SVD (Local Mean Decomposition-Singular Value Decomposition) and extreme learning machine |
| US20190042922A1 (en) * | 2018-06-29 | 2019-02-07 | Kamlesh Pillai | Deep neural network architecture using piecewise linear approximation |
| CN110569816A (en) * | 2019-09-12 | 2019-12-13 | 中国石油大学(华东) | Traffic sign classification algorithm based on new activation function |
-
2020
- 2020-05-20 CN CN202010431133.XA patent/CN111767981B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101404071A (en) * | 2008-11-07 | 2009-04-08 | 湖南大学 | Electronic circuit fault diagnosis neural network method based on grouping particle swarm algorithm |
| CN104408302A (en) * | 2014-11-19 | 2015-03-11 | 北京航空航天大学 | Bearing variable-condition fault diagnosis method based on LMD-SVD (Local Mean Decomposition-Singular Value Decomposition) and extreme learning machine |
| US20190042922A1 (en) * | 2018-06-29 | 2019-02-07 | Kamlesh Pillai | Deep neural network architecture using piecewise linear approximation |
| CN110569816A (en) * | 2019-09-12 | 2019-12-13 | 中国石油大学(华东) | Traffic sign classification algorithm based on new activation function |
Non-Patent Citations (2)
| Title |
|---|
| YAN LIU等: "《FPGA Implementation of a BPSK 1D-CNN Demodulator》", 《 APPLIED SCIENCES》 * |
| 杨景明等: "《基于FPGA的BP神经网络硬件实现及改进 》", 《计算机工程与设计》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022156475A1 (en) * | 2021-01-20 | 2022-07-28 | 华为技术有限公司 | Neural network model training method and apparatus, and data processing method and apparatus |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111767981B (en) | 2023-12-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112101530B (en) | Neural network training method, device, equipment and storage medium | |
| CN110942248B (en) | Training method and device for transaction wind control network and transaction risk detection method | |
| KR20060118598A (en) | Reference data optimization learning method and pattern recognition system | |
| CN112818588B (en) | Optimal power flow calculation method, device and storage medium of power system | |
| CN112990958B (en) | Data processing method, device, storage medium and computer equipment | |
| CN115935442A (en) | Block chain performance optimization method based on multi-agent deep reinforcement learning | |
| CN114781248A (en) | Off-line reinforcement learning method and device based on state offset correction | |
| CN111767981A (en) | Approximate calculation method of Mish activation function | |
| Lv et al. | Integrated double estimator architecture for reinforcement learning | |
| CN114998679A (en) | Online training method, device and equipment for deep learning model and storage medium | |
| CN118301002A (en) | Edge DPU computing force collaborative optimization unloading method for demand self-adaptive prediction | |
| CN109388784A (en) | Minimum entropy Density Estimator device generation method, device and computer readable storage medium | |
| CN111124439B (en) | Intelligent dynamic unloading algorithm with cloud edge cooperation | |
| CN112818986A (en) | Text detection method, device, system and medium based on depth relation reasoning | |
| CN117115551A (en) | Image detection model training method and device, electronic equipment and storage medium | |
| CN111462015A (en) | Map track processing method and device, electronic equipment and storage medium | |
| CN108599976B (en) | Network quality evaluation model construction method and device and computer equipment | |
| CN115423149A (en) | Incremental iterative clustering method for energy internet load prediction and noise level estimation | |
| CN107124159A (en) | A kind of implementation method of the particle filter based on adaptive KLD cassette lengths | |
| Guo et al. | On the first passage g-mean-variance optimality for discounted continuous-time Markov decision processes | |
| CN106651466A (en) | Website user-based information prediction method | |
| CN115564055A (en) | Asynchronous joint learning training method and device, computer equipment and storage medium | |
| Groeneboom | Nonparametric estimation of the incubation time distribution | |
| CN110705437A (en) | Face key point detection method and system based on dynamic cascade regression | |
| CN114065913A (en) | Model quantization method and device and terminal equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |