189 results about "Session state" patented technology

The invention provides an agent collaborative question-answering method and system based on context awareness and authority control, and belongs to the technical field of artificial intelligence. The method comprises the following steps: acquiring user query, and triggering a user context manager to generate a session state user context object containing a permission label and preference; the central scheduler performs task decomposition based on the query intention and the context, dynamically selects an intelligent agent through a multi-dimensional scoring model and binds a knowledge source; the authority-aware agent execution engine executes triple authority verification and intermediate result desensitization processing in the whole process, and supports agent fault tolerance and audit log recording. The method realizes unification of knowledge security and access flexibility, supports multi-scene adaptive question and answer, and has good expandability and fault-tolerant capability.

The invention discloses a soft network system capable of expanding a transmission mode and an implementation method, and the system comprises an equipment resource mapping management unit which automatically generates a unique channel code based on a native hardware interface and establishes a multi-dimensional mapping relation; the heterogeneous interface data acquisition unit is used for monitoring various hardware interfaces in a kernel space in parallel and realizing zero-copy data transmission through a three-buffer lockless annular buffer area; the transmission abstraction and dynamic expansion unit realizes transmission mode modularization in a plug-in form and supports atomic hot plug during operation; the protocol analysis and service processing unit is used for realizing unified scheduling and fault isolation in a single process through a dynamic link mechanism; and the cross-layer monitoring and self-healing unit realizes automatic switching based on a session state machine and a redundant channel. According to the method, the code complexity is reduced from O (N * M) to O (N + M), and the time lt is extended; fault switching is carried out at 1s; and the hardware cost is reduced by more than 40%, so that the method is particularly suitable for power and environment monitoring scenes with more than 1000 devices and more than 15 protocols, and has high creativity.

The invention discloses a self-adaptive threat blocking method and system oriented to industrial control protocol deep analysis, and the method comprises the steps: constructing an industrial control protocol multi-dimensional analysis model, extracting an operation code semantic rule, a parameter structure constraint and session state hopping features based on protocol reverse engineering, generating a protocol syntax tree, and blocking a threat from the protocol syntax tree. Wherein the operation code semantic rule is matched with the service primitive mapping of an abstract communication service interface in the IEC 61850 standard; based on the protocol syntax tree, semantic-level analysis is carried out on real-time communication flow, dynamic indexes are extracted, the dynamic indexes comprise a process value mutation gradient, an instruction sequence logic conflict degree and a session state abnormal jump rate scale, and a three-dimensional threat feature vector is generated. According to the method, hidden attacks are accurately recognized through industrial control protocol semantic analysis and multi-dimensional dynamic blocking, so that the defense action and the production stability can be collaboratively optimized.

The invention relates to a resource elastic scaling decision-making method, system and device and a medium. The method comprises the following steps: collecting real-time operation data of a security service node, and performing multi-dimensional security index analysis according to the real-time operation data to obtain a portrait data packet; predicting the security service weight value to obtain a prediction result, performing dynamic error compensation on the prediction result to generate a corrected weight prediction value, and generating a control instruction based on the corrected weight prediction value and the active session state; and when the instruction is a migration instruction, analyzing a session state snapshot of the instruction, calling a preset kernel state locking function to lock a memory session block of a source node, obtaining incremental state change data to generate a migration snapshot packet, and performing block verification injection operation on a target node. According to the method, by integrating multi-dimensional safety index analysis, prediction error compensation and stateful transition verification mechanisms, the accuracy and response efficiency of resource elastic scaling decision making are improved, and the continuity of stateful service transition and the consistency of safety strategies are enhanced.

The invention relates to the technical field of encrypted communication, and discloses an intelligent analysis method and system based on multi-source data, and the method comprises the steps: obtaining multi-dimensional metadata of an encrypted communication flow, and generating a feature vector set; calculating a state transition probability of a communication session by using a Markov chain model, dividing the communication session according to a time window and mapping the communication session to a discrete state, and generating a session state evolution sequence; analyzing a session state evolution sequence based on a hierarchical topic model, mapping the state sequence to a behavior topic through potential Dirichlet allocation, mapping the behavior topic to an intention category through a hierarchical Dirichlet process, and outputting a communication intention category and a confidence score; according to the method, the limitation that static feature analysis cannot reflect the complete life cycle of the session is overcome, the problem of semantic gaps caused by limited metadata information dimensions is solved, and the accuracy and robustness of intention recognition are improved.

The invention discloses a cross-platform session state reconstruction method and system based on event traceability, and relates to the field of cross-platform customer service integration, and the method comprises the steps: introducing a measurable business influence weight for a cross-platform field event through an event formatting mode based on business semantics, and maintaining a cross-platform event sequence in combination with a vector clock; according to the method, the real key events in the business meaning are adaptively selected to generate a state abstract through the time sequence accumulation calculation of the event weights, and snapshot is not carried out based on fixed time or a fixed number, so that the data transmission quantity and the calculation overhead required by state reconstruction are remarkably reduced; in the recovery stage of the client, only the abstract and the increment event after the abstract need to be transmitted, so that the client can still quickly reestablish the state at lower cost in a weak network and low bandwidth environment. Therefore, differential compression and recovery of a cross-platform event traceability semantic level are realized, and session reconstruction efficiency and resource utilization efficiency in a customer service large-scale multi-platform scene are greatly improved.

The invention relates to the field of man-machine interaction, and particularly provides a task processing method, device, equipment and system, a medium and a product. The method comprises the following steps: in response to a first session instruction of a user for a target type task, executing the target type task; the target type task belongs to a task type capable of receiving supplementary information input by a user during task execution; in the task execution period of the target type task, responding to a received second session instruction input by the user, and executing a corresponding task operation according to the intention of the second session instruction; wherein the intention is used for indicating to input supplementary information of the target type task or initiate a new task. Therefore, on the premise of not interrupting the original target type task, the new task instruction of the user can be responded in parallel or the supplementary input of the user to the target type task can be processed, and the problems of interaction blockage and context chaos caused by the exclusive session state under the existing architecture are solved.

The invention provides an edge session seamless migration method, system and device and a storage medium, which are suitable for real-time session scenes such as remote desktops, screen streaming media, interaction control and the like. The method comprises the following steps: generating a migration token based on trigger conditions such as load, link quality and proximity, executing snapshot and incremental log copying of a session state, establishing a parallel communication channel in a target relay, and completing NAT penetration and quality detection; calculating a sequence number and time offset between a source relay and a target relay, constructing a repetition suppression and ACK mapping mechanism, realizing mirror image drainage in a parallel window, and completing main path switching after a QoE condition is satisfied; and rollback control and key context synchronization are supported, and the stability and security of the migration process are ensured. The scheme has the advantages of low time delay, no interruption, high reliability and the like, and is suitable for various edge deployment and complex network environments.

The invention provides a user-defined role switching method and device for an AI voice module, and relates to the technical field of data processing, and the method comprises the steps: receiving a role switching instruction inputted by a user, carrying out the identification binding of the original data of a current session with a first role before switching, and recording the current dialogue state of the first role; according to a role switching instruction input by a user, determining a switching target as a second role, and extracting and archiving historical context information in the first role session data set; screening the context transfer data set in combination with the identity information, the permission level and the theme tag of the second role, and extracting historical context content suitable for the second role; updating the current session state with the second role; when the user switches back to the first role again, calling the stored current dialogue state and the dialogue data set of the first role; the autonomy and accuracy of the user-defined role switching method are improved.

The invention relates to the field of electric digital data processing and artificial intelligence session systems, and discloses a dynamic intention recognition and multi-agent seamless switching session system and method based on a large language model, and the system comprises an LLM central routing module, an agent session state management module and a cross-platform agent adapter pool. The intelligent agent session state management module defines a main context heap in a physical memory and manages a memory address of a private context block, the LLM central routing module maintains an active context pointer, and session switching is achieved by redirecting the pointer to a target memory address. Therefore, the problems of system delay and resources caused by high-frequency I / O (input / output) are avoided.

The embodiment of the invention provides a user session processing method and device, electronic equipment and a storage medium, and relates to the technical field of user session processing. The method comprises the following steps: acquiring a complexity score of a current user request and a current session state; according to the complexity score of the current user request and the current session state, performing state migration processing on the current user request according to a preset state migration rule, and determining a target session state of the current user request; according to the method, the target back-end response model is determined based on the target session state, and the current user request is routed to the target back-end response model for response processing, that is, intelligent routing of the user request is realized by combining the complexity score with a dynamic migration mechanism of the session state, and the service cost is effectively controlled while the response quality is guaranteed.

The invention relates to the field of load balancing, and particularly discloses a national standard signaling load balancing system and a national standard signaling load balancing method based on elb and redis, which are characterized in that registration and session state data of all equipment are stripped from a signaling server and are converted into a high-availability shared cache for centralized persistence, so that the signaling server becomes a stateless computing node. When the load balancer detects that any signaling server fails, the load balancer accurately and smoothly redirects the subsequent signaling of the affected device to another healthy takeover server in the cluster by using the orderliness of the consistent Hash algorithm. As all the servers maintain the complete mirror image of the global device state in the local memory in advance, the takeover server can immediately identify the device and seamlessly recover the session. In this way, the problems of session loss and large-scale re-registration storm caused by server faults are solved.

The embodiment of the invention provides a multi-round dialogue complexity assessment method and device, electronic equipment and a computer readable storage medium, and relates to the technical field of artificial intelligence. According to the method, by fusing the trend of the session state, the semantic turning of the abstract and the correlation of the historical dialogues, the evaluation accuracy of the real complexity of the request in multiple rounds of dialogues is improved, and misjudgment caused by only depending on the original text length or keywords is avoided; and meanwhile, the context length transmitted to the large language model is effectively compressed by using the abstract containing intention transformation and the related dialogue, so that the Token cost is directly saved, the calculation overhead brought by redundant context transmission is reduced, and a more efficient and more intelligent model routing decision is supported. Accurate context compression and routing decision reduce the average duration of request processing, and the user experience can be improved.

The invention relates to the technical field of crossing of information security and artificial intelligence, in particular to a system and method for accurately fusing system login permission based on GNN risk assessment, comprising the following steps: S1, constructing an initial permission topological graph comprising user nodes, token nodes and resource nodes, and setting risk weights for edges in the graph according to historical abnormal access behaviors; loading a plurality of sets of preset permission strategy libraries and preheating map neural network model parameters; s2, executing first-layer cryptographic signature verification of the token on the received access request to confirm the validity of the token; analyzing a session state identifier carried in the token after the signature verification is passed, and switching to a corresponding permission policy library at the current thread-level atom according to the identifier; according to the method, key technologies such as dynamic authority topology analysis based on the graph neural network, session-level real-time fusing and zero-copy auditing are introduced, so that millisecond-level accurate detection and isolation of the user access risk are realized.

The application provides a business data security protection method and system for digital enterprise management, which comprises the following steps: obtaining an access request sequence initiated by a target user on a business operation terminal, generating a dynamic access control policy vector according to a historical behavior track associated with an access operation identifier, and analyzing the dynamic access control policy vector through a policy decoder to generate a real-time access permission instruction; based on the real-time access permission instruction, homomorphic encryption mapping is performed on sensitive data segments in the request context information to generate a ciphertext transmission channel, and the ciphertext transmission channel is synchronized with a security parameter of a cross-department joint analysis model; a federal learning framework is called to perform multi-modal feature fusion on a real-time operation flow of the business operation terminal, an abnormal operation confidence score is generated, and when the score exceeds a dynamic threshold, a cross-level interception protocol is triggered and the current session state is rolled back to a safe baseline version. The application can improve the accuracy and response timeliness of anomaly detection.

According to an aspect, a method includes receiving, by a language model, session state information, generated by a browser application, about a computer session of a user, receiving, by the language model, a user-generated query including a natural language description about a browser tab, generating, by the language model, a textual response about the browser tab using contextual data, the contextual data including the session state information, and providing the textual response on a user device.

This invention discloses a highly interactive SIP honeypot system based on phased interactive control, comprising: a message receiving module for receiving external SIP requests; a SIP interactive control module supporting multiple interaction modes and generating corresponding responses according to the current configuration mode; a session state machine module for maintaining the SIP session protocol state; a phase switching control module for maintaining phase state variables and dynamically selecting the interaction mode according to preset trigger conditions; and a recording and statistics module for recording request behavior characteristics, interaction logs, and switching logs. This invention achieves runtime schedulable management of interactive capabilities, enabling low-resource-consumption recording of scanning behavior and deep trapping of advanced attacks and collection of complete attack chain data, providing reliable data support for the detection of encrypted malicious traffic.

The invention provides a Spring Session performance optimization method based on partition cache, and relates to the technical field of session management and cache optimization in a distributed system, and the method comprises the steps: storing a mapping relation between a session ID and a final access timestamp by designing concurrent hash mapping caches C1 and C2 which are alternately used, and identifying an active cache by using a current cache label; configuring a timed task to periodically scan the cache, judging a session state according to the critical zone time, and dynamically updating or cleaning cache data; a session object obtained by a thread local variable cache Redis is introduced in Http request processing, so that repeated access is avoided; and a Spring session filter and a warehouse are rewritten to reconstruct session access logic, so that the dependence on Redis is effectively reduced, and the concurrency performance and response efficiency of the system are improved.

The invention relates to a quick game making method and a game making cloud platform, and the method comprises the steps: binding received natural language demand description data, sketch track data and real-time voice streams into an associated data object which can be stored for a long time, and distributing a global unique ID and a version updating record identifier for the data object; according to the response emergency degree of the instruction in the data object, distributing a differentiated data transmission channel; performing time point matching on the voice features, the sketch structure features and the text semantic features of the same data object through a space-time synchronization engine, and outputting a unified instruction vector; calling a pre-trained game content generation model, and converting the unified instruction vector into game elements which can be independently stored; and monitoring the modification operation of a player on the game elements, extracting modification difference data, training and generating a model optimization module, and updating a version record. According to the method, the effects of cross-session state inheritance of the NPC / scene, dynamic preemptive scheduling of computing power resources and reverse correction of the AI model by the player operation data are realized.

The application relates to the field of data processing and discloses an e-commerce advertisement pushing management method and system based on data processing, which comprises the following steps: capturing the page rendering, sliding rhythm and content switching track in a session on the terminal side, constructing an interaction rhythm signal for representing the current session interaction intensity according to the operation time interval fluctuation amplitude and trigger frequency change; adaptively adjusting the time window boundary according to the user interaction burst density and terminal reporting delay, generating a session state offset parameter for putting prediction control; constructing a putting prediction model in the current session time window based on the session state offset parameter, generating a putting mismatch trigger signal for triggering the advertisement pushing strategy correction; binding the corresponding user instance and advertisement content instance, generating dynamic putting control parameters for the current session; and performing real-time correction on the advertisement pushing sorting and parameter strategy according to the dynamic putting control parameters. The application has the advantage of improving the real-time adaptability of advertisement putting.

The application discloses a device control method, device, storage medium and computer program product, relates to the computer technical field, and the method comprises the following steps: when receiving a hibernation instruction, a trusted execution environment is created, and a dynamic session state plaintext is exported from a trusted platform module to the trusted execution environment; the dynamic session state plaintext is encrypted by using a pre-generated hibernate binding key in the trusted execution environment to obtain dynamic session state ciphertext, and a check value corresponding to the dynamic session state ciphertext is calculated; a hibernate image is created based on the dynamic session state ciphertext and the corresponding check value in the trusted execution environment, and the hibernate image is stored in a nonvolatile storage medium; and a hibernate operation is triggered. The application improves the wake-up efficiency of the device from the hibernate state.

The invention relates to the technical field of financial transactions, in particular to a transaction front-end processor failover method and device, and the method comprises the following steps: a cluster management service monitors the state of a front node on a ZK cluster; when the front-end node session is in a free state, the cluster management service changes the state of the session in the online transaction front-end processor through message driving; the transaction API is reconnected to the online transaction front-end processor, and a subscription instruction a is sent; the online transaction front-end processor refreshes the push stream file according to the subscription instruction a; when the front-end node session is in a failure state, the transaction API is connected to an online transaction front-end processor, and a subscription instruction b is sent; and recovering the transaction of the context, and performing data refreshing of the transaction private stream. According to the method, the transaction API is used for registering a plurality of front-end processors, and a complete session state is maintained, so that the consistency of transaction contexts and subscription instructions of the transaction API are ensured, the transaction contexts are quickly recovered, and the average fault-free time and the reliability of a transaction system are improved.

The application relates to an adaptive multi-channel remote gesture interaction method and system, which comprises the following steps: generating a working mode identifier according to a sensor available state; collecting sensor original data streams and solving to generate standardized gesture vectors; inputting the vectors into a trajectory prediction model trained based on historical data to generate prediction trajectory data; combining current channel bandwidth characteristics and the prediction data to perform frame rate adaptation and motion smoothing processing on the gesture vectors to generate optimized instruction data packets; acquiring a user operation event, performing hierarchical local feedback while sending the instruction data packets; when it is monitored that a network packet loss rate exceeds a threshold value, generating a session state snapshot and synchronizing the snapshot to a backup channel to complete a hot switching; analyzing instructions and snapshot rendering cursor trajectories at a remote end, and generating an updated session state snapshot based on trajectory behavior characteristics; and storing the snapshot into a trajectory tracking database for subsequent state recovery. The application can improve user remote interaction experience in a complex scene.

The invention relates to a link switching method and system, computer equipment and a storage medium. The link switching method and system are realized on the basis of virtualized Socket layers deployed on a main control side and a module side of an OpenCPU heterogeneous system. The method comprises the following steps: in response to a migration triggering condition, initiating a migration request of a network session between a master control side and a module side; generating a session state of the network session at a source end currently holding the network session, and transmitting the session state subjected to serialization processing to a target end; and at the target end, based on the received session state, recovering and taking over the network session. According to the application, the virtualization Socket layers are deployed on the master control side and the module side, and serialization and synchronization of session states of the master control side and the module side are realized by the virtualization Socket layers, so that communication continuity of network sessions is ensured, and the problem that lossless link switching between the master control side and the module cannot be realized is solved.

The invention is suitable for the technical field of network security, and provides a computer network security assessment method and system, and the method comprises the steps: collecting the connection establishment request data with a timestamp of a target network, and extracting the multi-dimensional dynamic characteristics of a connection establishment process; constructing and maintaining a dynamic normal connection establishment behavior baseline in combination with network context dimensions such as a network topology position; comparing the real-time multi-dimensional dynamic features with corresponding baselines to generate abnormal deviation indexes, and associating session state data to distinguish malicious connection establishment behaviors and legal service fluctuations; and performing security situation assessment and scoring on each logic unit of the network according to the result, and positioning a security event and generating early warning information in combination with network topology information. According to the method, multi-dimensional feature mining and refined baseline matching of network connection behaviors are realized, network abnormity and service fluctuation are accurately distinguished, the accuracy of security assessment is improved, accurate positioning and early warning of security events can be realized, and safe operation of the network is effectively guaranteed.

Techniques are disclosed for mitigating lost state data for a communication session by a network device performing session-based routing. For example, in response to determining a first network device has not received packets of a reverse packet flow of a session from a second network device, the first network device sends, to the second network device, a first packet including first metadata comprising a first control message querying a health of the session. The first network device receives, from the second network device, a second control message requesting second metadata for reestablishing the session, the second control message sent in response to the first control message. The first network device sends a second packet of the forward packet flow including the second metadata for reestablishing the session to the second network device. The second network device recovers lost state data for the session using the second metadata.

The present disclosure provides a system (108) and a method for re-establishing a Protocol Data Unit (PDU) session during a failure in a network (106). The system (108) receives a service accept message from a network (106). The system (108) detects an absence of an Information Element (IE) comprising a PDU session status in the service accept message. In response to the detection, the system (108) frequently sends a service request message to the network (106) for a predetermined number of attempts and intervals. The system (108) detects that the network (106) fails to append the IE comprising the PDU session status in the service accept message within the predetermined number of attempts. Based on the detection, the system (108) performs a PDU session release and re-establishes the PDU session with the network (106).

The present application relates to the technical field of key management, and particularly relates to a key management method for unmanned aerial vehicles, a medium and equipment, a symmetric session key uniquely associated with a flight session identifier is cooperatively generated by no less than a first preset number threshold of symmetric key generation nodes, which breaks away from single node dependence, and realizes precise matching of the key and the flight session life cycle; the key state and the flight session state are synchronized by continuously maintaining the key effective state and real-time detecting the triggering condition of the revocation condition during the flight session operation, and the security risks caused by the key being valid when the session is abnormal are avoided; after the revocation condition is triggered, the key is marked as invalid when the recommended revocation node number reaches a second preset number threshold, which avoids abnormal revocation of the key caused by single node misoperation or malicious behavior, disperses the node load, solves the performance bottleneck of centralized key management, and improves the processing efficiency and scalability.

The application relates to an interactive stage switching method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: determining a conversation state of a current interactive stage based on multi-round interactive information, and judging whether the current interactive stage is at a switching node according to a change trend of the conversation state; if not, converting the conversation state into an interactive prompt in the current interactive stage; after displaying the interactive prompt in the current interactive stage, updating the multi-round interactive information according to obtained current interactive information; if yes, switching the current interactive stage into a next interactive stage according to an interactive stage sequence; the interactive prompts generated by the current interactive stage and the next interactive stage are different. The method can trace a control process.