Time induction mechanism implementation method and device, equipment and storage medium

Abstract

The invention provides a time induction mechanism implementation method and device, equipment and a storage medium. Relates to the technical field of artificial intelligence. The method comprises the following steps: constructing a global buffer, a time sequence association processing module, a causal relationship access module and a main memory; storing a causal relationship by using a main memory; acquiring input event information, and temporarily storing the event information by using a global buffer; performing preliminary event induction on the event information by using a time sequence association processing module to obtain association strength of the event information; and the event information of which the association strength exceeds a set threshold value is extracted from the time sequence association processing module by utilizing the causal relationship access module, the event information is stored in the main memory, the related causal relationship associated with the event information in the causal relationship is called from the main memory during expectation, and the expected causal relationship is generated and fed to the main memory. According to the method, the problems of efficiency and accuracy of a time induction mechanism in a general artificial intelligence system are solved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular, to a method, apparatus, device, and storage medium for implementing a time induction mechanism. Background Art

[0002] With the continuous advancement of general artificial intelligence, the current expectations for the reasoning ability and environmental adaptability of intelligent systems are increasing day by day. As a key link for intelligent systems to process event time sequences and causal inferences, the time induction mechanism plays a central role in cognitive activities such as intelligent systems' response to dynamic changes and planning future actions.

[0003] Currently, the time induction mechanism in artificial intelligence systems mainly focuses on in-depth exploration of issues related to event time sequence relationship processing and the accuracy and efficiency of predicting future events. Existing research mainly focuses on the following aspects:

[0004] 1. Data collection and causal relationship construction: By collecting time series data of event occurrences, analyzing the order of events, constructing causal relationship statements, and integrating them into the knowledge structure. The construction of causal relationships is usually based on the temporal coherence of events, their occurrence frequencies, and their potential impact on expected future events.

[0005] 2. Inductive reasoning algorithms: From simple temporal associations to complex logical reasoning, a variety of algorithms are applied to the time induction process. Techniques such as rule-based reasoning, probabilistic graphical models, and experience-driven reasoning are widely used in this field, aiming to enhance the reliability and adaptability of the time induction mechanism so that it can better cope with complex and changing environments.

[0006] 3. Mechanism evaluation and application: By simulating scenario tests or actual case applications, evaluate the expected accuracy and resource utilization efficiency of the time induction mechanism. The research focuses on improving the fit between expected events and actual occurrences, the rational use of computing resources, and enhancing the prediction ability of the system in a dynamic environment.

[0007] An existing method for implementing a temporal induction mechanism focuses on the direct storage and retrieval of simple causal relationships, which makes it difficult to balance the accurate judgment of causal relationships and the efficient use of storage structures. When faced with complex and diverse event sequences, this method may lead to confusion in knowledge storage, making it difficult to quickly obtain correct causal relationships, while incorrect causal relationships frequently interfere with system reasoning. Because it only uses a simple table based on the order of event occurrence for causal storage, when an expected search is required, the entire table will be traversed, resulting in extremely low efficiency and a large consumption of computing resources. At the same time, in the process of learning causal relationships, as long as the events present a temporal order, the causal relationship is mechanically determined to exist and stored in the memory. This rough processing method makes it very easy for the system to learn incorrect causal relationships. This seriously hinders the performance improvement and application expansion of artificial intelligence systems in temporal induction.

[0008] Disadvantages of existing technology:

[0009] (1) Rigid induction methods: Most existing solutions in the field of machine learning are only applicable to simple scenarios with single data features, and usually rely on applying past cases and established algorithms to process stable data. Once in a complex and dynamic environment, these solutions are unable to effectively explore the potential causal logic between events, and are unable to flexibly adjust the induction method according to environmental changes, and thus are unable to achieve efficient sorting and accurate prediction of time.

[0010] (2) Indiscriminate expectations: In the time induction phase, traditional general artificial intelligence systems store causal relationships and simple event sequence judgment criteria in a table format. When performing expected operations, they will indiscriminately retrieve all relevant causal statements and are completely unable to distinguish their importance and degree of correlation, which in turn leads to indiscriminate expectations. This makes it difficult for the system to focus on key causal relationships for effective expectations, seriously affecting the accuracy and reliability of expectations.

[0011] (3) Waste of computing resources: In the expected operation of the above-mentioned traditional system, due to the indiscriminate retrieval of all causal statements, a large amount of computing resources are used to process information that may have little or no relevance to the current expectations, resulting in serious waste of computing resources. This not only reduces the operating efficiency of the system, but also undermines the rationality of resource utilization, making it difficult for the system to respond quickly when dealing with complex dynamic environments due to unreasonable consumption of resources.

[0012] (4) Deficiency in false causal learning: The time induction mechanism of traditional general artificial intelligence systems determines and stores causal relationships based on the simple sequence of events, lacking a rigorous verification and screening process. When faced with complex event sequences and variable environmental factors, this rough processing method is extremely likely to lead the system to learn false causal relationships. Once false causal connections accumulate in the system, it will seriously interfere with subsequent reasoning and prediction processes, significantly reducing the prediction accuracy of the system in complex dynamic scenarios and severely limiting the reliability and practicality of the system.

[0013] These drawbacks severely restrict the application effects of existing technologies in many complex real-world scenarios such as intelligent traffic prediction and complex system fault diagnosis. Therefore, there is an urgent need for a new method for implementing a time induction mechanism based on general artificial intelligence, which can deeply explore the potential causal logic between events in a complex and variable environment, enable the general artificial intelligence system to autonomously, real-time and accurately judge the sequence and causal relationship of events, quickly adapt to environmental changes through flexible induction strategies, achieve efficient sorting and accurate prediction of time relationships, significantly improve the time induction efficiency of the general artificial intelligence system, promote its advancement towards a higher level of intelligence, and thus effectively meet the urgent need for accurate time analysis and event prediction in complex real-world scenarios. Summary of the Invention

[0014] This application provides a method, device, equipment and storage medium for implementing a time induction mechanism, aiming to overcome the problems of efficiency and accuracy in the time induction mechanism of general artificial intelligence systems, and constructs an innovative time induction architecture for general artificial intelligence. By reshaping the causal relationship storage structure and refining the information processing flow, it enables the general artificial intelligence system to achieve efficient and accurate time induction and event prediction based on the temporal correlation and dynamic causal logic of events in a complex and variable environment without the need for a large amount of additional computing resources. Although current research on general artificial intelligence involves time induction, the original mechanism is difficult to make effective predictions quickly and accurately when making predictions, while the innovative architecture of this application can precisely fill this key gap, strongly promoting the performance leap and intelligent evolution of general artificial intelligence in practical application scenarios.

[0015] In a first aspect, this application provides a method for implementing a time induction mechanism, including:

[0016] Construct a global buffer, a temporal correlation processing module, a causal relationship access and storage module, and a main memory;

[0017] Use the main memory to store causal relationships; wherein, the way of using the main memory to store causal relationships includes: setting each event and its corresponding causal relationship as nodes in a network, and connecting each node with weights;

[0018] Obtain the input event information and temporarily store the event information using the global buffer;

[0019] Use the timing correlation processing module to extract event information from the global buffer and perform preliminary event induction on the extracted event information to obtain the correlation strength of the event information;

[0020] Use the causal relationship access module to extract the event information with a correlation strength exceeding the set threshold from the timing correlation processing module, store it in the main memory, and retrieve the relevant causal relationships associated with the event information in the causal relationship from the main memory when making predictions, generate the expected causal relationship, and feed it to the main memory;

[0021] When the main memory obtains the expected causal relationship, update the currently stored causal relationship based on the expected causal relationship.

[0022] In a possible design, the method of performing preliminary event induction on the extracted event information includes:

[0023] When the event information is extracted for the first time, assign an initial weight as the correlation strength of the event information according to the time interval between multiple event information. As the event sequence is observed multiple times, increase the weight between the corresponding event information according to the number of associations between multiple event information.

[0024] In a possible design, use the causal relationship access module to retrieve the causal relationship associated with the event information from the main memory, and select the causal relationship with the highest weight to generate the expected causal relationship.

[0025] In a possible design, when the main memory obtains the expected causal relationship, update the currently stored causal relationship based on the expected causal relationship through the following method:

[0026] Match and verify the generated expected causal relationship with the actual causal relationship of the event:

[0027] If the expected causal relationship is inconsistent with the actual causal relationship, reduce the weight of the causal relationship;

[0028] If the expected causal relationship is consistent with the actual causal relationship, increase the weight of the causal relationship.

[0029] In a possible design, the method further includes: constructing an emotion module; before using the temporal association processing module to extract event information from the global buffer, using the emotion module to extract event information from the global buffer and determining the emotional intensity of the event information, and when the emotional intensity of the event information exceeds a set emotional intensity threshold, using the temporal association processing module to extract event information from the global buffer and directly feeding the event information to the main memory, and the main memory updates the causal relationship in response to the input event information.

[0030] In a second aspect, the present application provides a device for implementing a time induction mechanism, the device includes:

[0031] A main memory configured to store causal relationships and, after the expected causal relationship matches the actual causal relationship, update the currently stored causal relationship according to the difference between the two. When storing the causal relationship, the main memory sets each event and its corresponding causal relationship as nodes in the network, and the nodes are interconnected by weights;

[0032] A global buffer configured to obtain input event information and temporarily store the event information;

[0033] A temporal association processing module configured to extract event information from the global buffer and perform preliminary event induction on the extracted event information to obtain the association strength of the event information;

[0034] A causal relationship access module configured to extract event information with an association strength exceeding a set threshold from the temporal association processing module, store it in the main memory, and at the same time retrieve the causal relationship associated with the event information from the main memory to generate an expected causal relationship;

[0035] Match and verify the generated expected causal relationship with the actual causal relationship of the event, and update the causal relationship stored in the main memory according to the matching result.

[0036] In a possible design, the device further includes an emotion module configured to extract event information from the global buffer and determine the emotional intensity of the event information;

[0037] The temporal association processing module is further configured to, when the emotional intensity of the event information exceeds a set emotional intensity threshold, extract event information from the global buffer and directly feed the event information to the main memory;

[0038] The main memory is further configured to update the causal relationship in response to the input event information.

[0039] In a third aspect, an embodiment of the present application provides an electronic device, including: at least one processor and a memory; the memory stores computer-executable instructions; the at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the time induction mechanism implementation method described in the first aspect above and various possible designs of the first aspect.

[0040] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the time induction mechanism implementation method described in the first aspect above and various possible designs of the first aspect is implemented.

[0041] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program. When the computer program is executed by a processor, the time induction mechanism implementation method described in the first aspect above and various possible designs of the first aspect is implemented.

[0042] The time induction mechanism implementation method, device, equipment and storage medium provided by the present application have at least the following beneficial effects:

[0043] The present application endows the general artificial intelligence system with powerful time induction ability and demonstrates significant advantages in many aspects.

[0044] In terms of accuracy, the present invention breaks through the shackles of the traditional time induction mechanism, draws on advanced models in human brain cognitive science, and innovatively reshapes the causal storage and processing process. By constructing a complex causal relationship storage architecture similar to a neural network, accurately representing events and their causal connections with nodes and weights, the system can efficiently screen and judge according to the size of the weights and the degree of correlation when facing a large amount of event information, greatly improving the accuracy of causal judgment, effectively reducing the probability of learning wrong causal relationships due to traditional simple judgment criteria, and thus providing a solid and reliable basis for the subsequent reasoning and decision-making of the system.

[0045] In terms of the rationality of resource utilization, the present application abandons the traditional undifferentiated retrieval and processing methods and adopts an intelligent resource allocation strategy. During the time induction and causal reasoning processes, it can accurately identify key events and important causal paths, focus computing resources on core tasks, and avoid unnecessary waste and dispersion of resources. At the same time, for low-priority events and causal associations that still have potential value, appropriate attention and processing can also be given when the system resources are abundant, achieving an effective balance of resource allocation among different levels of tasks, ensuring the maximized utilization of the time induction function of the system under limited resource conditions, and comprehensively improving the comprehensive efficiency and application value of the general artificial intelligence system in time induction. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The accompanying drawings herein are incorporated into and constitute a part of this specification, showing embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application.

[0047] Figure 1 It is a flowchart of a method for implementing a time induction mechanism provided by an embodiment of the present application;

[0048] Figure 2 It is a schematic diagram of a time induction architecture for general artificial intelligence provided by an embodiment of the present application;

[0049] Figure 3 It is a schematic diagram of a time induction architecture for general artificial intelligence including an emotion module provided by an embodiment of the present application;

[0050] Figure 4 It is a schematic diagram of the structure of a device for implementing a time induction mechanism provided by an embodiment of the present application;

[0051] Figure 5 It is a schematic diagram of the structure of a device for implementing a time induction mechanism including an emotion module provided by an embodiment of the present application.

[0052] Through the above accompanying drawings, specific embodiments of the present application have been shown, and there will be more detailed descriptions hereinafter. These drawings and written descriptions are not intended to limit the scope of the concept of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] Here, exemplary embodiments will be described in detail, and the examples are shown in the accompanying drawings. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0054] In the technical solution of the present application, the processing of the collection, storage, use, processing, transmission, provision, and disclosure of financial data, user data, and other information involved all comply with the provisions of relevant laws and regulations and do not violate public order and good customs.

[0055] It should be noted that in the embodiments of the present application, some industry-existing solutions such as certain software, components, models, etc. may be mentioned. They should be considered exemplary, and their purpose is only to illustrate the feasibility in the implementation of the technical solution of the present application, but it does not mean that the applicant has already or necessarily used this solution.

[0056] The technical solution of the present application and how the technical solution of the present application solves the above technical problems will be described in detail below with specific embodiments. The following several specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of the present application will be described below with reference to the accompanying drawings.

[0057] In view of the fact that in the prior art, machine learning and traditional general artificial intelligence systems can only rely on simple rules and established models to process data in temporal induction, it is difficult to deeply understand the causal logic between events, and it is impossible to achieve a flexible and efficient temporal induction process. The embodiments of the present application provide a method for implementing a temporal induction mechanism, which is applicable to event time series analysis and causal reasoning in general artificial intelligence systems, especially in the fields of intelligent traffic prediction, complex system fault diagnosis, intelligent regulation of smart home environments, etc. It can significantly improve the accuracy of mining causal relationships between events in a complex dynamic environment, and enhance the response speed and processing ability of the system to changes in time series, thereby improving the overall intelligence level and application efficiency of general artificial intelligence systems.

[0058] Generally speaking, the core of the method for implementing the temporal induction mechanism in the embodiments of the present application lies in endowing the general artificial intelligence system with a powerful and accurate prediction ability. By repeatedly observing events, the system can collect more information about the details of event occurrence and the performance in different situations. In multiple observations, the system can gradually filter out the situations that occur accidentally in chronological order but actually have no causal relationship, so as to more accurately identify the event pairs that truly have a causal connection. In a complex and changeable environment, the system can, based on the rich experience accumulated, achieve accurate prediction of the development trends of various events.

[0059] Figure 1 It is a flowchart of a method for implementing a temporal induction mechanism provided by an embodiment of the present application. As Figure 1 shown, the method for implementing the temporal induction mechanism provided by the embodiments of the present application includes the following steps S10 to S60.

[0060] S10: Construct a global buffer, a temporal correlation processing module, a causal relationship access module, and a main memory.

[0061] Exemplarily, as Figure 2As shown in the figure, it is a schematic diagram of a time induction architecture for general artificial intelligence provided by an embodiment of the present application. The time induction architecture for general artificial intelligence includes a global buffer 100, a temporal correlation processing module 200, a causal relationship access module 300, and a main memory 400. Among them, the role of the global buffer 100 is to cache event information. The event information can be input into the global buffer 100 through multiple channels. The information input through multiple channels includes Narsese, knowledge, perception information, and other inputs. In the global buffer 100, the event information and non-event information are inserted into the priority list according to the priority. The priority can be determined, for example, according to the chronological order of the inputs. The temporal correlation processing module 200 is used to obtain event information from the temporal correlation processing module 200 and form event sequences, combine multiple event sequences and perform event induction, and transfer the event sequences to the causal relationship access module 300 after reaching a given threshold. The causal relationship access module 300 stores causal relationships and combines the causal relationships stored in the main memory 400 to generate expectations, obtaining expected causal relationships. The main memory 400 obtains the corresponding event information from the global buffer 100 and updates the stored causal relationships based on the expected causal relationships. In the main memory 400, the causal relationships are stored in the form of packets (probability priority queues).

[0062] S20: Use the main memory to store causal relationships; among them, the way of using the main memory to store causal relationships includes: setting each event and its corresponding causal relationship as nodes in the network, and connecting the nodes through weights.

[0063] In this embodiment, considering that in traditional general artificial intelligence systems, the storage form of event causal relationships is extremely simple. It mainly adopts a table-like structure, taking the events that occur first as the main body and recording the causal relationships between the related events. However, this method has serious defects. In the face of a large number of complex events, it is easy to lead to undifferentiated expectations, that is, the system cannot accurately judge which causal relationships are more relevant and important, resulting in resource waste and expectation deviation.

[0064] In contrast, the mode of storing causal relationships in the human brain is much more complex and efficient, existing in the form of a complex network. Inspired by this, this embodiment draws on the connectionist model in the field of cognitive science and deeply imitates the structure of the human brain neural network. Specifically, each event and its corresponding causal relationship are set as nodes in the network. These nodes do not exist in isolation, and they are closely connected through weights. The size of the weights intuitively reflects the connection strength between the nodes.

[0065] When the system needs to make predictions, it no longer blindly considers all possible causal paths, but instead intelligently selects the path with the highest weight. This process highly mimics the way humans make judgments in daily life, based on the importance and frequency of past experiences. In this way, the system can effectively avoid indiscriminate predictions, quickly locate the most valuable information among a vast number of causal relationships, greatly improving the quality of causal relationship storage and the accuracy during retrieval, laying a solid foundation for the efficient execution of the general artificial intelligence system in temporal induction tasks.

[0066] S30: Obtain the input event information and temporarily store the event information using the global buffer.

[0067] In traditional general artificial intelligence systems, as long as events are observed to occur in chronological order, a causal relationship where the earlier event causes the later event will be formed, making it easy to learn incorrect causal relationships.

[0068] However, when the human brain processes causal relationships, the following process occurs: When external information floods in, first, the sensory organs capture various information, such as visual images seen and sounds heard. These sensory information then enter the sensory memory, which is similar to a short-term "buffer zone" where the information is only temporarily stored without any in-depth processing, merely maintaining the original state of the information. After that, the sensory memory distributes the information to various specialized areas of the brain. Among them, event information related to time is transmitted to the hippocampus. The key role of the hippocampus in the brain is to perform preliminary temporal induction and association on the perceived events. It is like an "association builder" that simply connects some events or information in chronological order. During this process, as the number of times observing the association of the same event increases, the neural connections between these events will gradually be strengthened, just like "leaving an imprint" in the brain. After preliminary processing by the hippocampus, this information is transferred to other brain areas such as the prefrontal cortex. The prefrontal cortex is equivalent to the "decision-making commander" of the brain, where a more in-depth analysis and judgment of the causal relationship between events are carried out. After finally determining the causal relationship, it is stored in long-term memory for subsequent retrieval at any time.

[0069] Therefore, in this embodiment, the global buffer is similar to the sensory memory in the human brain, serving only as an information temporary storage area without any processing. Its role is to ensure the integrity and originality of the information in the initial reception stage, avoiding misinterpreting or processing the information before sufficient analysis, and laying a foundation for accurately judging causal relationships later.

[0070] In some embodiments, the method further includes: constructing an emotion module. Such as Figure 3As shown, it is a schematic diagram of a temporal induction architecture for general artificial intelligence including an emotion module 500 provided by an embodiment of the present application. Before extracting event information from the global buffer 100 using the temporal correlation processing module 200 in step S40, the emotion module 500 extracts event information from the global buffer 100 and determines the emotional intensity of the event information. When the emotional intensity of the event information exceeds a set emotional intensity threshold, the temporal correlation processing module is used to extract event information from the global buffer 100 and directly feed the event information to the main memory 400. In response to the input event information, the main memory 400 updates the causal relationship.

[0071] When emotions are involved, especially in a state of high emotional intensity, such as experiencing extreme danger and generating fear, or obtaining a major surprise and triggering excitement, the information processing flow will change significantly. At this time, when the hippocampus processes event information related to this emotion, the speed and degree of neural connection strengthening far exceed normal. Different from the normal situation, this kind of information full of strong emotions will not be transferred to the prefrontal cortex for in-depth analysis and judgment as usual. This is because under the drive of strong emotions, the brain needs to quickly respond to this information with high importance or urgency. These information will directly enter the long-term memory storage area from the hippocampus quickly. For example, when a person suddenly faces a life-threatening emergency, the fear emotion will cause the brain to quickly store the information related to this dangerous scene directly into the long-term memory, so that when encountering a similar situation later, an instinctive reaction can be made quickly. This processing method greatly shortens the information processing path, greatly improves the timeliness of information processing, and ensures that individuals can make quick responses at critical moments.

[0072] Inspired by the above, consider that in the case of high emotional intensity, the situation is different. When the emotion module 500 gives feedback of high emotional intensity, it means that the event has a high degree of importance or urgency. At this time, the weight of the relevant information is very high at the beginning, and there is no need to be transmitted to the causal relationship access module 300 after reaching a high weight through multiple cumulative associations in the temporal association processing module 200 as in the normal case. Instead, it directly enters the main memory 400 from the temporal association processing module 200 to ensure that the system can quickly respond to and store such key information, greatly improving the timeliness of information processing. Exemplarily, the emotion module 500 can process the event information based on the existing emotion classification model to obtain the emotion classification probability as the emotional intensity. When the emotion classification probability is greater than the set emotional intensity threshold, it is determined as a high emotional intensity situation. For example, in the application scenario of fault diagnosis, when high-intensity words such as "major", "urgent", and "serious" are included in the input event information, it is determined that the current event information belongs to the high emotional intensity situation, and the event information can be directly entered into the main memory 400 through the fast process. The main memory 400 can update the causal relationship based on the input event information in the following way: directly add a node to represent the input event information, determine the relationship between the newly added node and other original nodes, and determine the weight for connecting other nodes.

[0073] S40: Use the temporal association processing module to extract event information from the global buffer and perform preliminary event induction on the extracted event information to obtain the association strength of the event information.

[0074] In this embodiment, the temporal association processing module simulates the progressive association strengthening function of the parahippocampal gyrus. In traditional general artificial intelligence systems, incorrect causal relationships are often learned by simply determining causality based on the order of events. This module is responsible for preliminary temporal induction and establishing the initial association between events. By introducing a weight mechanism, as the number of associations increases, the weight gradually increases, and at the same time, a time decay mechanism is added to simulate the human forgetting process. This means that if an association only occurs accidentally, its weight will decrease over time, avoiding being solidified as long-term memory due to accidental associations, thus effectively reducing the probability of learning incorrect causal relationships. Only when the association strength reaches a specific threshold will the information be transmitted to the next module, further improving the quality of information screening and processing, and ensuring that the causal relationships contained in the information entering the subsequent processing have high reliability.

[0075] In some embodiments, when performing preliminary event induction on the extracted event information, when an event first enters, an initial weight is first assigned based on the temporal proximity between events, and the number of associations is not considered at this time. For example, when events A, B, and C enter the global buffer in sequence, the temporal correlation processing module will assign weights to event pairs according to the time interval: Since the time interval between events A and B is the shortest, a relatively high initial weight is assigned, such as 0.4; the time interval between events B and C is slightly longer, and the weight assigned is 0.35; while the time interval between events A and C is the longest, and the weight assigned is 0.3. However, only when the weight of an event pair reaches a preset threshold (e.g., 0.6) will it be passed to the causal relationship access module for further processing. As the system observes the event sequence multiple times, the number of associations between events A and B increases, and their weights will also increase accordingly. When the weight reaches or exceeds 0.6, the causal relationship between events A and B will be passed to the causal relationship access module. And the weight will gradually decrease over time.

[0076] S50: Use the causal relationship access module to extract event information with an association strength exceeding the set threshold from the temporal correlation processing module, store it in the main memory, and retrieve the relevant causal relationships associated with the event information in the causal relationship from the main memory when making predictions, generate the expected causal relationship, and feed it back to the main memory.

[0077] In this embodiment, the causal relationship access module receives strong association information from the temporal correlation processor and stores it in the main memory. When external information enters the system, it is responsible for calling relevant causal relationships from the main memory to generate expectations. As the interface between the main memory and the processing system, it undertakes the tasks of storing and retrieving causal relationships. Through a rigorous information reception and storage process, as well as an accurate call mechanism, this module ensures that the system can quickly and accurately call effective causal information in the face of a complex and changing environment, avoiding incorrect time induction and event expectations caused by calling the wrong causal relationship, and thus achieving accurate time induction and event expectations.

[0078] Exemplarily, use the causal relationship access module to extract event information with an association strength greater than the set threshold from the temporal correlation processing module, and store its causal relationship in the main memory. When making predictions in the causal relationship access module, call the causal relationship stored in the main memory, and select the relationship with the highest weight to generate the expected causal relationship. This expected causal relationship needs to be verified by matching with the actual causal relationship. According to the matching result, update the main memory based on the difference between the expected causal relationship and the actual causal relationship: if the two are inconsistent, reduce the weight of this causal relationship; if they are consistent, increase its weight.

[0079] In some embodiments, the set correlation strength threshold is determined according to the weights in the causal relationship. For example, the average value of the weights in the causal relationship is used as the set correlation strength threshold.

[0080] S60: When the expected causal relationship is obtained in the main memory, update the currently stored causal relationship based on the expected causal relationship.

[0081] Exemplarily, when the similarities and differences between the expected causal relationship and the actual causal relationship are obtained in the main memory, update the currently stored causal relationship based on this relationship. If the expected causal relationship is consistent with the actual causal relationship, increase the weight of the currently stored causal relationship; otherwise, decrease it.

[0082] In this embodiment, the expected causal relationship is the relationship between the newly added event information and the events existing in the original causal relationship. The newly added event information is also represented in the form of nodes, and the nodes related to it are connected by weights. Therefore, the method for updating the currently stored causal relationship based on the comparison between the expected causal relationship and the actual causal relationship is as follows: If the result of the expected causal relationship is inconsistent with the actual result, directly add a new node in the currently stored causal relationship network to represent the new event information, and decrease the weight of this expected causal relationship; if the result of the expected causal relationship is consistent with the actual result, increase the weight of this expected causal relationship.

[0083] In summary, the core of the implementation method of this temporal induction mechanism lies in two aspects: causal relationship storage and processing flow. In terms of causal relationship storage, drawing on the connectionist model in the field of cognitive science and imitating the structure of the human brain neural network, events and their causal relationships are set as nodes, and the connection strength of the nodes is reflected by weights. In the processing flow, information first enters the global buffer for temporary storage, and then undergoes preliminary induction and correlation strengthening by the temporal correlation processing module. The information that reaches the threshold enters the causal relationship access module for storage and retrieval. Under high emotional intensity, the information transmission path is shortened, and it directly enters the memory from the temporal correlation processing module. The entire structure cooperates closely to improve the system's temporal induction and anticipation capabilities in a complex and changing environment.

[0084] For causal relationship storage, this embodiment adopts the idea of the connectionist model in cognitive science and proposes to store the causal relationships between events in the form of a network. The events and their causal relationships are regarded as nodes in the network at the same time, and weights are used to represent the connection strength between events. This storage method mimics the neural network structure in the human brain, similar to the connections between neurons and synaptic plasticity. Through dynamic weight adjustment, the strengthening and weakening mechanisms of causal relationships in the human brain are simulated, making the storage of causal relationships more efficient and adaptable. The path with the highest weight is selected for prediction. This method can better simulate the process of humans making judgments based on the importance and frequency of past experiences when facing multiple possibilities, thus avoiding indiscriminate prediction.

[0085] For the processing flow, through multi-module collaborative processing, the causal relationship processing is divided into a global buffer, a temporal association processing module, and a causal relationship access module. The global buffer ensures the integrity and originality of the initial information; the temporal association processing module simulates the hippocampus to achieve preliminary time induction and association strengthening, and screens reliable causal relationships through a weight mechanism and a time decay mechanism; the causal relationship access module is responsible for accurately storing and efficiently retrieving causal information. This way of multi-module collaborative operation, different from the traditional single or simple combined processing mode, is the key to achieving efficient time induction.

[0086] This embodiment of the present application also provides a device for implementing a time induction mechanism, as Figure 4 shown. The device for implementing the time induction mechanism includes a global buffer 100, a temporal association processing module 200, a causal relationship access module 300, and a main memory 400;

[0087] The main memory 400 is configured to store causal relationships and, after the predicted causal relationship matches the actual causal relationship, update the currently stored causal relationship according to the difference between the two. When storing causal relationships, the main memory sets each event and its corresponding causal relationship as nodes in the network, and the nodes are interconnected through weights;

[0088] The global buffer 100 is configured to obtain the input event information and temporarily store the event information;

[0089] The temporal association processing module 200 is configured to extract event information from the global buffer and perform preliminary event induction on the extracted event information to obtain the association strength of the event information;

[0090] The causal relationship access module 300 is configured to extract event information whose correlation strength exceeds a set threshold from the timing association processing module and store it in the main memory, and at the same time retrieve the causal relationship associated with the event information from the main memory to generate an expected causal relationship; match and verify the generated expected causal relationship with the actual causal relationship of the event, and update the causal relationship stored in the main memory according to the matching result.

[0091] In some embodiments, Figure 5 As shown, the apparatus further includes an emotion module 500, and the emotion module 500 is configured to extract event information from the global buffer 100 and determine the emotion intensity of the event information;

[0092] The temporal association processing module 200 is further configured to extract event information from the global buffer 100 and directly feed the event information to the main memory 400 when the emotion intensity of the event information exceeds a set emotion intensity threshold;

[0093] The main memory 400 is further configured to update the causal relationship in response to input event information.

[0094] An embodiment of the present application provides an electronic device, which may include: a processor and a memory, wherein the processor and the memory may communicate with each other; illustratively, the processor and the memory communicate with each other via a communication bus.

[0095] The processor executes the computer execution instructions stored in the memory, so that the processor executes the scheme in the above embodiment. The processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit ASIC, a field programmable gate array FPGA or other programmable logic devices, discrete gates or transistor logic devices, and discrete hardware components.

[0096] The communication bus can be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The system bus can be divided into an address bus, a data bus, a control bus, etc. The transceiver is used to implement communication between the database access device and other computers (such as clients, read-write libraries, and read-only libraries). The memory may include Random Access Memory (RAM) and may also include non-volatile memory.

[0097] The electronic device provided by the embodiments of the present application can be the terminal device in the above embodiments.

[0098] The embodiments of the present application also provide a computer-readable storage medium storing computer instructions, which, when running on a computer, cause the computer to execute the technical solutions of the time induction mechanism implementation method in the above embodiments.

[0099] The embodiments of the present application also provide a computer program product including a computer program stored in a computer-readable storage medium. At least one processor can read the computer program from the computer-readable storage medium, and when the at least one processor executes the computer program, the technical solutions of the time induction mechanism implementation method in the above embodiments can be realized.

[0100] In several embodiments provided by the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of modules is only a logical function division. In actual implementation, there may be other division methods. For example, multiple modules can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of devices or modules can be in electrical, mechanical, or other forms.

[0101] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to implement the solutions of this embodiment.

[0102] In addition, in each embodiment of the present application, the functional modules can be integrated into one processing unit, or each module can exist physically alone, or two or more modules can be integrated into one unit. The unit formed by the above modules can be implemented in the form of hardware, or in the form of a combination of hardware and software functional units.

[0103] The integrated modules implemented in the form of software functional modules can be stored in a computer-readable storage medium. The above software functional modules are stored in a storage medium, including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to execute some steps of the methods in each embodiment of the present application.

[0104] It should be understood that the above processor can be a central processing unit (Central Processing Unit, abbreviated as CPU), and can also be other general-purpose processors, digital signal processors (Digital Signal Processor, abbreviated as DSP), application specific integrated circuits (Application Specific Integrated Circuit, abbreviated as ASIC), etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in combination with the invention can be directly embodied as being executed and completed by a hardware processor, or by a combination of hardware and software modules in the processor.

[0105] The memory may include high-speed RAM memory, and may also include non-volatile storage NVM, such as at least one disk memory, and can also be a USB flash drive, a mobile hard disk, a read-only memory, a magnetic disk, or an optical disc, etc.

[0106] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc.

[0107] The above storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The storage medium can be any available medium accessible by a general-purpose or special-purpose computer.

[0108] An exemplary storage medium is coupled to the processor, enabling the processor to read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can be located in an application specific integrated circuit (ASIC). Of course, the processor and the storage medium can also exist as discrete components in an electronic control unit or a master control device.

[0109] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps including the above method embodiments; and the foregoing storage medium includes various media that can store program codes, such as ROM, RAM, magnetic disk or optical disk.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for implementing a time summarization mechanism, characterized in that: The method comprises: Construct global buffer, temporal correlation processing module, causal relationship access module and main memory; The main memory is used to store the causal relationship; wherein the main memory is used to store the causal relationship in a manner including: setting each event and its corresponding causal relationship as a node in a network, and each node is connected by a weight; Acquire input event information, and use the global buffer to temporarily store the event information; Utilizing the time series association processing module, extracting event information from the global buffer, and performing preliminary event summarization on the extracted event information to obtain the association strength of the event information; The causal relationship access module is used to extract event information whose correlation strength exceeds the set threshold from the time series correlation processing module and store it in the main memory. When making an expectation, the relevant causal relationship associated with the event information in the causal relationship is retrieved from the main memory, and the expected causal relationship is generated and fed to the main memory; When the main memory obtains the expected causal relationship, the currently stored causal relationship is updated based on the expected causal relationship.

2. The method for implementing the time summarization mechanism according to claim 1, characterized in that: Methods for performing preliminary event summarization on the extracted event information include: When event information is extracted for the first time, an initial weight is assigned as the association strength of the event information according to the time interval between multiple event information. With multiple observations of the event sequence, the weights between the corresponding event information are increased according to the number of associations between the multiple event information.

3. The method for implementing the time summarization mechanism according to claim 1, characterized in that: The causal relationship access module is used to retrieve the causal relationships associated with the event information from the main memory, and the causal relationship with the highest weight is selected to generate the expected causal relationship.

4. The method for implementing the time summarization mechanism according to claim 1, characterized in that: When the main memory obtains the expected causal relationship, the currently stored causal relationship is updated based on the expected causal relationship by the following method: Verify the generated expected causal relationship by matching the actual causal relationship of the event: If the expected causal relationship is inconsistent with the actual causal relationship, the weight of the causal relationship is reduced; If the expected causal relationship is consistent with the actual causal relationship, the weight of the causal relationship is increased.

5. The method for implementing the time summarization mechanism according to claim 1, characterized in that: The method also includes: constructing an emotion module; before using the timing association processing module to extract event information from the global buffer, using the emotion module to extract event information from the global buffer and determine the emotion intensity of the event information; when the emotion intensity of the event information exceeds a set emotion intensity threshold, using the timing association processing module to extract event information from the global buffer and directly feed the event information to the main memory, and the main memory updates the causal relationship in response to the input event information.

6. A device for implementing a time summarization mechanism, characterized in that: The device comprises: a main memory, wherein the main memory is configured to store causal relationships, and after the expected causal relationships match the actual causal relationships, update the currently stored causal relationships according to the difference between the two, and when storing the causal relationships, the main memory sets each event and its corresponding causal relationship as a node in a network, and each node is connected to each other through weights; A global buffer, wherein the global buffer is configured to obtain input event information and temporarily store the event information; A time series correlation processing module, wherein the time series correlation processing module is configured to extract event information from the global buffer, and perform preliminary event summarization on the extracted event information to obtain correlation strength of the event information; A causal relationship access module, wherein the causal relationship access module is configured to extract event information whose correlation strength exceeds a set threshold from the temporal correlation processing module, store the event information in the main memory, and retrieve the causal relationship associated with the event information from the main memory to generate an expected causal relationship; The generated expected causal relationship is matched and verified with the actual causal relationship of the event, and the causal relationship stored in the main memory is updated according to the matching result.

7. The device for implementing the time summarization mechanism according to claim 6, characterized in that: The apparatus further comprises an emotion module configured to extract event information from the global buffer and determine an emotion intensity of the event information; The temporal association processing module is further configured to extract event information from the global buffer and directly feed the event information to the main memory when the emotion intensity of the event information exceeds a set emotion intensity threshold; The main memory is further configured to update the causal relationship in response to input event information.

8. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the time induction mechanism implementation method according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the time induction mechanism implementation method according to any one of claims 1 to 5 when executed by a processor.

10. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for implementing a time induction mechanism according to any one of claims 1 to 5 is implemented.

Images