CN111185010A - System and method for constructing landlord card-playing program by using pulse neural network - Google Patents
System and method for constructing landlord card-playing program by using pulse neural network Download PDFInfo
- Publication number
- CN111185010A CN111185010A CN201911352866.8A CN201911352866A CN111185010A CN 111185010 A CN111185010 A CN 111185010A CN 201911352866 A CN201911352866 A CN 201911352866A CN 111185010 A CN111185010 A CN 111185010A
- Authority
- CN
- China
- Prior art keywords
- playing
- card
- neurons
- layer
- signals
- 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
Images
Classifications
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63F—CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR
- A63F13/00—Video games, i.e. games using an electronically generated display having two or more dimensions
- A63F13/60—Generating or modifying game content before or while executing the game program, e.g. authoring tools specially adapted for game development or game-integrated level editor
- A63F13/67—Generating or modifying game content before or while executing the game program, e.g. authoring tools specially adapted for game development or game-integrated level editor adaptively or by learning from player actions, e.g. skill level adjustment or by storing successful combat sequences for re-use
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63F—CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR
- A63F13/00—Video games, i.e. games using an electronically generated display having two or more dimensions
- A63F13/80—Special adaptations for executing a specific game genre or game mode
- A63F13/822—Strategy games; Role-playing games
-
- 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
- G06N3/045—Combinations of networks
-
- 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
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
Abstract
A system for constructing a landlord card-playing program using a spiking neural network, comprising: the card type sensing module, the card-playing decision-making module and the card-playing module. The card type sensing module is composed of two groups of independent neurons, receives the code combination of the playing cards played by the previous family respectively to judge the types and sizes of the playing cards played by the current family, senses the playing cards held by the current family and counts possible playing methods. The card-playing decision-making module is composed of three groups of independent neurons, and each group corresponds to different scenes in the card game: actively playing cards, passively playing cards and friends in the former family, passively playing cards and enemies in the former family. The card-playing decision module obtains a possible card-playing combination of the house according to the pulse signal transmitted by the card type sensing module, transmits the pulse signal to the next card-playing module by using the optimal card-playing method, and finishes card-playing by the card-playing module. The invention applies the impulse neural network to the poker game of the poker game.
Description
Technical Field
The invention belongs to the technical field of artificial intelligence and neural networks, and particularly relates to a system and a method for constructing a landlord card-playing program by using a pulse neural network.
Background
In early electronic games, the implementation method of chess and card type robot players was mainly improved based on the classic algorithms such as backtracking method and genetic algorithm. Such a method relying on mechanization or statistical method often requires a large number of operations, and its effect cannot get rid of mechanization and has no learning ability. The rise of artificial intelligence and the improvement of computing power of hardware equipment enable realization of intelligent players by utilizing machine learning and neural networks to be realized, at present, the most advanced robot players are realized by adopting a second-generation neural network, and the principle is based on methods such as gradient descent, error back-propagation and the like. The method not only brings huge computation amount, but also has longer time consumption and lower precision, and the accuracy and the timeliness can not be ensured on the premise of energy conservation and general hardware computation.
The pulse neural network is a third-generation artificial neural network, fills the gap between neuroscience and machine learning by using a bionics principle, encodes information into membrane potential and pulse time delay of neurons, achieves high-efficiency and low-energy-consumption information transmission by using pulses, and is closer to the structure of a biological neural network. Currently, the impulse neural network has been widely applied to the problems of digital recognition, pattern recognition, and the like.
Disclosure of Invention
In order to overcome the above-mentioned shortcomings of the prior art, the present invention provides a system and method for constructing a ground-engaging card-playing program by using a neural network, which can greatly reduce the computational requirements and improve the processing speed and intelligence.
In order to achieve the purpose, the invention adopts the technical scheme that:
a system for constructing a card-playing program of a landlord by using a pulse neural network adopts a three-layer pulse neural network structure, and comprises the following steps:
the card type sensing module corresponds to a first layer of neurons, receives a playing card coding combination played by a last family, senses the characteristics of a playing card held by the own family, stimulates the first layer of neurons by playing card coding signals, and sends an inhibition signal or a stimulation signal to a second layer of neurons by the first layer of neurons;
the card-playing decision module is used for comprehensively inhibiting signals and exciting signals corresponding to a second layer of neurons to obtain all possible card-playing methods, deducing a sub-optimal card-playing method, and sending pulse signals to a third layer of neurons by controlling the pulse delay length so that the pulse signals reaching the third layer of neurons firstly are the sub-optimal card-playing method, namely the activated neurons represent the sub-optimal card-playing;
and the card-playing module corresponds to the third layer of neurons and receives the pulse signal of the card-playing decision module to finish the card-playing action.
Preferably, the neurons in the first layer are divided into two groups, each group comprises 54 neurons, and different pulse signals formed by coding 54 different playing cards are sensed correspondingly, wherein the neurons in group 2 are responsible for receiving information of playing cards played by a user, and transmitting inhibition signals to the neurons in the second layer, and all playing card methods which do not accord with game rules can receive the inhibition signals by the neurons in the second layer; the neurons of the group 1 are responsible for receiving information of playing cards held by a user, sending excitation signals to the neurons of the second layer which do not receive the inhibition signals, and sending pulse signals with certain time delay to the neurons of the third layer, wherein the corresponding neurons of the second layer can be activated when all card types of the playing cards which are played by the user can be pressed.
Preferably, the neurons in the second layer are divided into three groups, each group comprises 276 neurons and corresponds to 276 playing card combination modes, wherein the neurons in the group 1 correspond to a fighting strategy of 'passively playing cards and enemy to the last family', the neurons in the group 2 correspond to a fighting strategy of 'actively playing cards', and the neurons in the group 3 correspond to a fighting strategy of 'passively playing cards and friend to the last family'; when the former family is friends, the bigger the priority of the card-playing combination is higher, the corresponding neuron has shorter pulse time delay; when the cards are actively played, the higher the priority of the playing combination with the largest number of the played cards is, the shorter the pulse time delay of the corresponding neuron is; when the former family is enemy, the smaller the card-out combination has higher priority, and the corresponding neuron has the higher pulse delay.
Preferably, the neurons in the third layer are a group, and include 276 neurons, which correspond to 276 playing card combinations respectively.
The invention also provides a method for constructing a system for playing the card by the landlord based on the pulse neural network, which comprises the following steps:
1) stimulating a first layer of neurons of the pulse neural network by signals coded by the playing cards, and sending inhibition signals or excitation signals to a lower layer by the first layer of neurons;
2) and (3) comprehensively inhibiting signals and exciting signals to obtain all possible card-playing methods in the neurons of the second layer, and controlling the pulse delay length to ensure that the pulse signal which firstly reaches the neurons of the third layer is the best method for playing cards at the time, thereby finishing the card-playing action.
Further, the step 1) comprises the following steps:
step 1.1), inputting the playing card code played by the last family and the playing card code held by the family
Coding 54 different playing cards into different pulse signals, sending corresponding potential pulses to neurons which are in charge of processing the playing card signals of the upper family at a first layer of a neural network according to the playing cards of the upper family, and then sending corresponding potential pulses to the neurons which are in charge of processing the playing card signals of the upper family at the first layer according to the playing cards held by the family;
step 1.2), inhibiting neurons in the second layer part
The neurons of the first layer which are responsible for processing the playing signals of the previous house transmit inhibition signals to the neurons of the second layer, and the neurons of the corresponding second layer receive the inhibition signals according to the playing rules of the landlord and all playing methods which do not accord with the game rules;
step 1.3), exciting the neurons in the second layer part
And (3) sending an excitation signal to neurons of a second layer which are not inhibited by the step 1.2) by the neurons of the first layer which are responsible for processing the signals of the own playing cards, and according to the playing rules of the landlord, all card types which can be played by the player can be pressed, the corresponding neurons of the second layer can be activated, and a pulse signal with a certain time delay is sent to the neurons of the next layer.
Further, the step 2) comprises the following steps:
step 2.1), sending pulse signals to the third layer of neurons
The second layer of activated neurons are dealt according to different strategies according to different scenes, namely active card-playing, passive card-playing, friend-playing, passive card-playing and enemy-playing, when the friends are the friends of the family, the method for setting the larger number of the card-playing types corresponds to the shorter time delay of the pulse sent by the neurons, when the active card-playing is carried out, the method for setting the larger number of the card-playing types corresponds to the shorter time delay of the pulse sent by the neurons, and when the enemy is the enemy, the method for setting the smaller size of the card-playing types corresponds to the shorter time delay of the pulse sent by the neurons;
step 2.2), the third layer neuron decides the card and sends the card code
The third layer of neurons sequentially receive pulse signals with different time delays sent by the second layer, the signal which arrives at the first layer is the optimal card-playing method, and after the third layer of neurons send card-playing codes, the neurons which are played in the card-playing inspection layer of the own hand are frozen, and one card-playing action is finished;
step 2.3), cyclic operation
After the card-playing action of the step 2.2) is finished, entering a state of waiting for the previous card-playing till the previous card-playing is finished, returning to the step 1.1) to carry out the next round of card-playing action till the card game is finally divided into a winner and a loser.
Compared with the existing moving object speed identification system and method, the invention has the beneficial effects that:
1. the method can flexibly deal with the card-playing in different situations while observing the rules of the fighting-landlord game by constructing the three-layer impulse neural network, and compared with other technologies in the current period, the method has lower requirement on hardware computing power, has universality and higher efficiency.
2. The method utilizes a multi-stage neuron delay cascade structure, uses pulses to transmit information, and is more flexible in response, higher in speed and more energy-saving.
3. Compared with other technologies in the current period, the working principle of the method is closer to the functional principle of nerve cells, and the method has more advanced theoretical support and more development potential in the field of artificial intelligence.
Drawings
Fig. 1 is a schematic block diagram of the present invention.
FIG. 2 is a schematic diagram of the three-layer neural network structure of the present invention.
Figure 3 is a schematic diagram of a first layer of neurons receiving playing card signals from a parent and a parent, delivering inhibitory or positive potential pulses to a second layer of neurons.
Fig. 4 shows that the neurons in the second layer transmit pulse signals according to different groups corresponding to different priority rules, wherein each pulse signal has different time delays, and the pulse signals transmitted by the neurons closer to the front in the priority rules have shorter time delays.
Detailed Description
The embodiments of the present invention will be described in detail below with reference to the accompanying drawings and examples, which illustrate in detail the implementation of a framework for constructing a ground-fighting main card-playing program by using a neural network in an automatic card-playing situation.
The invention relates to a system framework for constructing a main card-playing program of a fighting field by using a pulse neural network, which comprises a card type sensing module for receiving a playing card code combination of a last-family playing card and sensing all the card codes of a family, a decision-making module for deducing the best card-playing in the current time and a card-playing module.
Referring to fig. 1, a system for constructing a ground-fighting main card-playing program by using a pulse neural network adopts a three-layer pulse neural network structure, and comprises:
the card type sensing module corresponds to a first layer of neurons, receives a playing card coding combination played by a last family, senses the characteristics of a playing card held by the own family, stimulates the first layer of neurons by playing card coding signals, and sends an inhibition signal or a stimulation signal to a second layer of neurons by the first layer of neurons; the card game machine comprises a hand viewing layer for viewing the last home playing cards and viewing the own-hold playing cards.
The card-playing decision module is used for comprehensively inhibiting signals and exciting signals corresponding to a second layer of neurons to obtain all possible card-playing methods, deducing a sub-optimal card-playing method, and sending pulse signals to a third layer of neurons by controlling the pulse delay length so that the pulse signals reaching the third layer of neurons firstly are the sub-optimal card-playing method, namely the activated neurons represent the sub-optimal card-playing; the card-playing decision-making layer structure is adopted, wherein the card-playing decision-making layer structure corresponds to three rules of active card-playing and passive card-playing (friend or enemy of the former family).
And the card-playing module corresponds to the third layer of neurons and receives the pulse signal of the card-playing decision module to finish the card-playing action.
Referring to fig. 1 and 4, the neurons in the first layer are divided into two groups, each group has 54 neurons, the neurons of group 1 constitute a hand viewing layer and are responsible for receiving information of playing cards held by a user, and the neurons of group 2 constitute a hand viewing layer and are responsible for receiving information of playing cards played by the user. The neurons in the second layer are divided into three groups, each group is provided with 276 neurons, and the neurons in the groups 1, 2 and 3 correspond to three fighting strategies, namely that the upper home is the landowner, the lower home is the landowner, and the upper home and the lower home are not the landowner respectively. When the former family is the ground, the bigger the card type, the higher the priority of the card-out combination, the shorter the pulse delay of the corresponding neuron. When the house is the landowner, the higher the priority of the card-playing combination with the larger number of the played cards is, the shorter the pulse delay of the corresponding neuron is, namely, the priority sequence is as follows, five-linked airplanes with wings, four-linked airplanes with four pairs of six-linked airplanes, four-linked airplanes with wings … … with three pairs of two, three pairs of one and three, pairs of one and single. When the upper home and the lower home are not the landowner, the smaller the card type, the higher the priority of the card-playing combination is, the shorter the pulse time delay of the corresponding neuron is, so as to cooperate with the teammate 'fighting the landowner'. There is only one group of neurons in the third layer, 276 neurons.
Referring to fig. 2 and 4, a schematic diagram of the neurons in the first layer receiving playing card signals from the upper house and the self-house and transmitting inhibitory pulses or positive potential pulses to the neurons in the second layer is shown in fig. 2, and as seen from fig. 2, firstly the neurons in the first layer group 2 are activated by playing cards played by the upper house, and according to the card type, it is necessary to exclude the playing card combination which does not meet the game rules from all the playing card possibilities (if the playing cards of the upper house are "3" and "3", the playing cards of the self-house cannot be "7"). Specifically, the operation in the neural network is that the activated neurons in the first layer group 2 send inhibition signals to the neurons in the second layer, and all the neurons corresponding to the card-playing modes which do not accord with the game rules are inhibited. The neurons of the first group of layers 1 are then activated by playing cards held by their own homes, sending positive potential pulse signals to the neurons of the second group of layers. Therefore, among the neurons that are not suppressed in the second layer, neurons that can correspond to a card-out combination that can be formed in the playing card held by the player on the pair of the player-type playing cards and the self-hold playing card are activated. Neurons # 1 and # 3 in the second packet 1 of figure 2 are activated.
Referring to fig. 3 and 4, the neurons in the second layer are activated and then transmit pulse signals with different time delays to the neurons in the third layer, for example, when the former is enemy and the playing cards played are "3" and "3", after signal processing and transmission from the first layer to the second layer, the neurons in the number 1 and the number 3 are both activated, and the playing cards played by the self-family are "4" and "4", or "a" and "a", and the priority of the neurons in the second layer set 1 is higher, so that the neurons in the number 1 can be suppressed to the maximum extent, therefore, the time delay y of the pulse sent by the neuron in the number 3 is smaller than the time delay x of the pulse sent by the neuron in the number 1, so that the pulse sent by the neuron in the number 3 reaches the neuron in the third layer first, and the neurons in the third layer receive the signal, i.e., the playing card, and then the most suitable card is played.
In conclusion, the invention forms an independent intelligent card-playing system framework of the poker game of the.
While the foregoing is directed to the preferred embodiment of the present invention, it is not intended that the invention be limited to the embodiment and the drawings disclosed herein. Equivalents and modifications may be made without departing from the spirit of the disclosure, which is to be considered as within the scope of the invention.
Claims (7)
1. A system for constructing a ground fighter card-playing program by using a pulse neural network is characterized in that a three-layer pulse neural network structure is adopted, and the system comprises:
the card type sensing module corresponds to a first layer of neurons, receives a playing card coding combination played by a last family, senses the characteristics of a playing card held by the own family, stimulates the first layer of neurons by playing card coding signals, and sends an inhibition signal or a stimulation signal to a second layer of neurons by the first layer of neurons;
the card-playing decision module is used for comprehensively inhibiting signals and exciting signals corresponding to a second layer of neurons to obtain all possible card-playing methods, deducing a sub-optimal card-playing method, and sending pulse signals to a third layer of neurons by controlling the pulse delay length so that the pulse signals reaching the third layer of neurons firstly are the sub-optimal card-playing method, namely the activated neurons represent the sub-optimal card-playing;
and the card-playing module corresponds to the third layer of neurons and receives the pulse signal of the card-playing decision module to finish the card-playing action.
2. The system for constructing the ground-fighting main card-playing program by using the impulse neural network as claimed in claim 1, wherein the neurons in the first layer are divided into two groups, each group comprises 54 neurons, and the neurons correspondingly sense different impulse signals formed by coding 54 different playing cards, wherein the neurons in group 2 are responsible for receiving information of the playing cards played by the last family and transmitting inhibition signals to the neurons in the second layer, and all the playing methods which do not accord with game rules can be used, and the neurons corresponding to the second layer can receive the inhibition signals; the neurons of the group 1 are responsible for receiving information of playing cards held by a user, sending excitation signals to the neurons of the second layer which do not receive the inhibition signals, and sending pulse signals with certain time delay to the neurons of the third layer, wherein the corresponding neurons of the second layer can be activated when all card types of the playing cards which are played by the user can be pressed.
3. The system for constructing the ground-fighting owner card-playing program by using the impulse neural network as claimed in claim 2, wherein the neurons in the second layer are divided into three groups, each group comprises 276 neurons and corresponds to a playing card combination mode of 276 ground-fighting owner playing cards, wherein the neurons in group 1 correspond to a fighting strategy of 'passive playing card and enemy' and the neurons in group 2 correspond to a fighting strategy of 'active playing card' and the neurons in group 3 correspond to a fighting strategy of 'passive playing card and friends' respectively; when the former family is friends, the bigger the priority of the card-playing combination is higher, the corresponding neuron has shorter pulse time delay; when the cards are actively played, the higher the priority of the playing combination with the largest number of the played cards is, the shorter the pulse time delay of the corresponding neuron is; when the former family is enemy, the lower the card type, the higher the priority of the card-out combination, and the shorter the pulse delay of the corresponding neuron.
4. The system of claim 3, wherein the third layer of neurons comprises 276 neurons corresponding to 276 playing card combinations respectively.
5. A method for constructing a system for playing a card-fighting main card game by using a pulse neural network according to claim 1, which comprises the following steps:
1) stimulating a first layer of neurons of the pulse neural network by signals coded by the playing cards, and sending inhibition signals or excitation signals to a lower layer by the first layer of neurons;
2) and (3) comprehensively inhibiting signals and exciting signals to obtain all possible card-playing methods in the neurons of the second layer, and controlling the pulse delay length to ensure that the pulse signal which firstly reaches the neurons of the third layer is the best method for playing cards at the time, thereby finishing the card-playing action.
6. The method according to claim 5, wherein the step 1) comprises the steps of:
step 1.1), inputting the playing card code played by the last family and the playing card code held by the family
Coding 54 different playing cards into different pulse signals, sending corresponding potential pulses to neurons which are in charge of processing the playing card signals of the upper family at a first layer of a neural network according to the playing cards of the upper family, and then sending corresponding potential pulses to the neurons which are in charge of processing the playing card signals of the upper family at the first layer according to the playing cards held by the family;
step 1.2), inhibiting neurons in the second layer part
The neurons of the first layer which are responsible for processing the playing signals of the previous house transmit inhibition signals to the neurons of the second layer, and the neurons of the corresponding second layer receive the inhibition signals according to the playing rules of the landlord and all playing methods which do not accord with the game rules;
step 1.3), exciting the neurons in the second layer part
And (3) sending an excitation signal to neurons of a second layer which are not inhibited by the step 1.2) by the neurons of the first layer which are responsible for processing the signals of the own playing cards, and according to the playing rules of the landlord, all card types which can be played by the player can be pressed, the corresponding neurons of the second layer can be activated, and a pulse signal with a certain time delay is sent to the neurons of the next layer.
7. The method according to claim 5 or 6, wherein the step 2) comprises the steps of:
step 2.1), sending pulse signals to the third layer of neurons
The second layer of activated neurons are dealt according to different strategies according to different scenes, namely active card-playing, passive card-playing, friend-playing, passive card-playing and enemy-playing, when the friends are the friends of the family, the method for setting the larger number of the card-playing types corresponds to the shorter time delay of the pulse sent by the neurons, when the active card-playing is carried out, the method for setting the larger number of the card-playing types corresponds to the shorter time delay of the pulse sent by the neurons, and when the enemy is the enemy, the method for setting the smaller size of the card-playing types corresponds to the shorter time delay of the pulse sent by the neurons;
step 2.2), the third layer neuron decides the card and sends the card code
The third layer of neurons sequentially receive pulse signals with different time delays sent by the second layer, the signal which arrives at the first layer is the optimal card-playing method, and after the third layer of neurons send card-playing codes, the neurons which are played in the card-playing inspection layer of the own hand are frozen, and one card-playing action is finished;
step 2.3), cyclic operation
After the card-playing action of the step 2.2) is finished, entering a state of waiting for the previous card-playing till the previous card-playing is finished, returning to the step 1.1) to carry out the next round of card-playing action till the card game is finally divided into a winner and a loser.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911352866.8A CN111185010B (en) | 2019-12-25 | 2019-12-25 | System and method for constructing landlord card-playing program by using pulse neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911352866.8A CN111185010B (en) | 2019-12-25 | 2019-12-25 | System and method for constructing landlord card-playing program by using pulse neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111185010A true CN111185010A (en) | 2020-05-22 |
| CN111185010B CN111185010B (en) | 2021-08-03 |
Family
ID=70684589
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911352866.8A Active CN111185010B (en) | 2019-12-25 | 2019-12-25 | System and method for constructing landlord card-playing program by using pulse neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111185010B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112774185A (en) * | 2021-02-09 | 2021-05-11 | 腾讯科技（深圳）有限公司 | Virtual card control method, device and equipment in card virtual scene |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106055339A (en) * | 2016-06-08 | 2016-10-26 | 天津联众逸动科技发展有限公司 | Method for determining card playing strategy of computer player in two-against-one game |
| CN109508789A (en) * | 2018-06-01 | 2019-03-22 | 北京信息科技大学 | Predict method, storage medium, processor and the equipment of hands |
| CN109621422A (en) * | 2018-11-26 | 2019-04-16 | 腾讯科技（深圳）有限公司 | Electronics chess and card decision model training method and device, strategy-generating method and device |
| CN110152302A (en) * | 2018-02-23 | 2019-08-23 | 北京信息科技大学 | The hands of cards game generate and clustering method, equipment and processor |
| CN110175619A (en) * | 2018-11-27 | 2019-08-27 | 深圳市腾讯信息技术有限公司 | Method, equipment and storage medium are determined based on the board group of playing a card of machine learning model |
| CN110404265A (en) * | 2019-07-25 | 2019-11-05 | 哈尔滨工业大学（深圳） | A kind of non-complete information machine game method of more people based on game final phase of a chess game online resolution, device, system and storage medium |
| CN110457534A (en) * | 2019-07-30 | 2019-11-15 | 深圳市腾讯网域计算机网络有限公司 | A kind of data processing method based on artificial intelligence, device, terminal and medium |
-
2019
- 2019-12-25 CN CN201911352866.8A patent/CN111185010B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106055339A (en) * | 2016-06-08 | 2016-10-26 | 天津联众逸动科技发展有限公司 | Method for determining card playing strategy of computer player in two-against-one game |
| CN110152302A (en) * | 2018-02-23 | 2019-08-23 | 北京信息科技大学 | The hands of cards game generate and clustering method, equipment and processor |
| CN109508789A (en) * | 2018-06-01 | 2019-03-22 | 北京信息科技大学 | Predict method, storage medium, processor and the equipment of hands |
| CN109621422A (en) * | 2018-11-26 | 2019-04-16 | 腾讯科技（深圳）有限公司 | Electronics chess and card decision model training method and device, strategy-generating method and device |
| CN110175619A (en) * | 2018-11-27 | 2019-08-27 | 深圳市腾讯信息技术有限公司 | Method, equipment and storage medium are determined based on the board group of playing a card of machine learning model |
| CN110404265A (en) * | 2019-07-25 | 2019-11-05 | 哈尔滨工业大学（深圳） | A kind of non-complete information machine game method of more people based on game final phase of a chess game online resolution, device, system and storage medium |
| CN110457534A (en) * | 2019-07-30 | 2019-11-15 | 深圳市腾讯网域计算机网络有限公司 | A kind of data processing method based on artificial intelligence, device, terminal and medium |
Non-Patent Citations (2)
| Title |
|---|
| XDX2CT1314: "斗地主智能（AI）算法", 《HTTPS://BLOG.CSDN.NET/XDX2CT1314/ARTICLE/DETAILS/18798263#COMMENTS_12440779》 * |
| 胡志根: "脉冲神经网络学习算法的研究及其应用", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112774185A (en) * | 2021-02-09 | 2021-05-11 | 腾讯科技（深圳）有限公司 | Virtual card control method, device and equipment in card virtual scene |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111185010B (en) | 2021-08-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110404264B (en) | Multi-person non-complete information game strategy solving method, device and system based on virtual self-game and storage medium | |
| Arabzad et al. | Football match results prediction using artificial neural networks; the case of Iran Pro League | |
| CN108629422A (en) | A kind of intelligent body learning method of knowledge based guidance-tactics perception | |
| CN110141867A (en) | A kind of game intelligence body training method and device | |
| CN108159705A (en) | Matching process, device, storage medium and the electronic device of object | |
| CN111185010B (en) | System and method for constructing landlord card-playing program by using pulse neural network | |
| CN111282267A (en) | Information processing method, information processing apparatus, information processing medium, and electronic device | |
| CN109453524A (en) | A kind of method of object matching, the method for model training and server | |
| CN111729300A (en) | Monte Carlo tree search and convolutional neural network based bucket owner strategy research method | |
| Nakamura et al. | Constructing a human-like agent for the werewolf game using a psychological model based multiple perspectives | |
| Rosenberg et al. | Artificial swarm intelligence vs vegas betting markets | |
| James et al. | A comparative analysis of simplification and complexification in the evolution of neural network topologies | |
| Suznjevic et al. | Application context based algorithm for player skill evaluation in MOBA games | |
| Kopparapu et al. | Hidden agenda: a social deduction game with diverse learned equilibria | |
| CN112270103B (en) | Cooperative strategy inversion identification method based on multi-agent game | |
| CN110263937B (en) | Data processing method, device and storage medium | |
| Nicolai et al. | No-limit texas hold'em poker agents created with evolutionary neural networks | |
| Zhang et al. | Tactical reward shaping: Bypassing reinforcement learning with strategy-based goals | |
| Yegnanarayana et al. | Neural networks for contract bridge bidding | |
| Conradie et al. | Training bao game-playing agents using coevolutionary particle swarm optimization | |
| Ebrahimi et al. | Dynamic difficulty adjustment in games by using an interactive self-organizing architecture | |
| Willcox et al. | Dense neural network used to amplify the forecasting accuracy of real-time human swarms | |
| Dong et al. | The algorithm of spiking neural network and application in Poker games | |
| Li et al. | Study on the play strategy of dou dizhu poker based on convolution neural network | |
| CN1991848A (en) | Computer aided cue and ball selecting system and method therefor |
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 |