250 results about "Interactive Learning" patented technology

The invention discloses a multi-mode instruction analysis system for a humanoid robot. Comprising a voice input module, a visual input module, a voiceprint feature extraction module, an object recognition and pose estimation module, a multi-modal alignment network based on a space-time attention mechanism, a scene semantic tree construction module, an instruction node mapping module, a confidence evaluation module and a decision module. According to the system, accurate alignment of voice and visual information is realized through a space-time attention mechanism, environment information is represented in combination with a scene semantic tree structure, and the instruction analysis accuracy is improved. And dynamically evaluating the confidence coefficient by adopting a fuzzy instruction backtracking algorithm, and if the confidence coefficient is lower than a threshold value, starting multi-round dialogue clarification to reduce misoperation. According to the method, multi-modal data are fused, the historical interaction learning ability is optimized, the understanding efficiency and interaction robustness of complex instructions are remarkably improved, the method is suitable for scenes such as family service and logistics storage, and the intelligent level of man-machine cooperation is enhanced.

The invention discloses a multi-modal fusion rumor detection method and system based on a dynamic graph convolutional neural network. According to the method, a dynamic feature graph of a language propagation path is constructed, and potential features in the language propagation process are extracted and analyzed by utilizing time sequence changes and key node relations between nodes in a propagation graph. A neural network is adopted to extract and enhance image data, text semantic features are extracted in combination with a text feature modeling network, text feature vectorization expression is achieved based on a BERT model, and rich semantic information is obtained. And a gating mechanism is introduced to dynamically adjust fusion weights of different modal features, and an information fusion strategy is optimized. A collaborative attention mechanism is further adopted for deep fusion, interactive learning of text, image and propagation path features is enhanced, and the relevance of cross-modal and time series data is improved. And finally, inputting the fused feature vectors into a classifier for accurate classification, thereby realizing accurate detection of the social media rumors. According to the method, the multi-modal features are effectively integrated, and the false information identification efficiency is remarkably improved.

The invention relates to the technical field of artificial intelligence education, in particular to an intelligent homework tutoring system and method based on multi-modal interaction and adaptive learning, and the system comprises a multi-modal input analysis module, a cognitive state dynamic evaluation module, an intelligent decision explanation engine, and a learning effect visualization closed-loop module. The method has the beneficial effects that a three-dimensional adaptive system of explanation granularity-presentation form-interaction frequency is used, and teaching strategies such as visual derivation, concept metaphor and the like can be automatically matched according to cognitive styles of students. Secondly, a'backtracking reinforcement-lateral expansion 'double-intervention mechanism is innovatively used, and the recurrence rate of similar errors is effectively reduced through error real-time detection and correlation knowledge contrast teaching. And finally, constructing a dynamic knowledge graph and an interactive learning report, realizing visual tracking of a learning path, and helping students to establish a systematic knowledge framework.

The invention relates to the technical field of teaching management, discloses an intelligent teaching system based on a knowledge graph and virtual simulation and a construction method, and aims to effectively solve the difficulty in integration and management of teaching resources and provide personalized teaching experience. The system comprises a knowledge graph construction module which is used for integrating and standardizing teaching resources and constructing a knowledge graph containing knowledge points and association relationships thereof; the virtual simulation teaching module is used for displaying knowledge points in a virtual simulation form by constructing a knowledge presentation model and an interactive learning model, and evaluating learning conditions of students; the intelligent teaching control module is used for constructing a teaching strategy model by adopting a reinforcement learning algorithm and dynamically adjusting a teaching instruction according to the learning progress and the interaction effect of the students; and the execution presentation module is responsible for receiving and executing the teaching instruction and presenting the teaching content in the virtual teaching environment. The system improves the pertinence and effectiveness of teaching, and stimulates the learning interest of students.

The invention discloses a small molecule ligand drug screening method and system based on affinity prediction, and belongs to the technical field of biological medicine. The invention aims to solve the technical problem of low drug screening precision caused by molecular expression limitation, geometric invariance deficiency and insufficient multi-modal information fusion when virtual drug screening is carried out by using protein-ligand affinity. Comprising the following steps: acquiring ligand and protein structure information, and preprocessing to obtain coordinates and a feature matrix of ligand / pocket / residue; performing comprehensive representation, multi-feature flow self-adaption, geometric algebraic multi-layer perception and feature alignment processing on the feature matrix to obtain corresponding feature space representation; performing cross attention fusion and multi-scale interactive learning processing on the feature space representation in sequence to obtain fusion features; inputting the fused features into a multi-scale interactive learning module, and outputting final features; and finally, predicting the binding affinity of the ligand and the protein according to the fusion characteristics to obtain a binding affinity value.

The invention belongs to the field of computer vision image signal processing, and relates to a multi-robot collaborative indoor scene semantic segmentation method and system based on multivariate interactive learning. The method comprises the steps of adopting an image feature encoder based on a double attention integration module to realize cross-modal feature fusion on an RGB image and a depth image; a dimension anisotropy attention enhancement module is adopted, and feature optimization is carried out through a multi-dimensional direction attention mechanism; a cross-task collaborative interaction module is adopted, and information is shared among different tasks by utilizing an interaction mechanism based on a feature level; in a multi-task heterogeneous decoding stage, each robot uses a decoder with a specific task to execute a specific task, and semantic segmentation, instance segmentation and scene classification are realized; and a joint optimization loss function is adopted to perform end-to-end joint training on multiple tasks, so that overall optimization and collaborative awareness performance improvement are realized. According to the method, the precision and robustness can be improved when complex indoor scene semantic segmentation and other visual tasks are executed.

The invention relates to the technical field of computer vision and fire monitoring, and provides a fire image detection and segmentation method based on an end-to-end unified framework and physical knowledge embedding. Aiming at the problems of computation redundancy, feature segmentation and strong dependence on visible light caused by traditional staged processing, the invention provides the following technical scheme: constructing an end-to-end network comprising an Officient Hybrid Ender encoder and a mask-dino decoder, and realizing global modeling and cross-scale feature fusion through a single-layer Transform; thermal imaging physical knowledge embedding is innovatively introduced, and three fusion modes of pixel-level addition, feature-level Embedding and interactive learning are adopted; a lightweight single-layer Transform architecture and a multi-task loss function are designed, and GIOU, Dice and a joint detection segmentation hybrid matching strategy are combined. According to the method, a bounding box and mask prediction are synchronously generated through a unified query mechanism, the adaptability of a low-illumination scene is enhanced by using thermal imaging data, and the training efficiency is improved by decoupling bounding box loss. According to the invention, the real-time performance, robustness and precision of fire monitoring are significantly improved in a complex scene.

The invention discloses a scenarized hierarchical management and control method for an education application in a data space, and the method comprises the steps: obtaining a multi-modal input content in a corresponding preset period after the starting of the current time in response to a starting instruction of the education application in a target data space; calling a corresponding analysis strategy according to the content type of the multi-modal input content; extracting target features from the multi-modal input content based on an analysis strategy; according to an education scene corresponding to the target data space, obtaining a matching result of the target feature and a target database; determining a corresponding data space risk level according to a matching result; and calling and executing an application management and control strategy corresponding to the data space risk level. According to the method, education application hierarchical management and control adaptive to different scenes can be provided, the management and control precision is high, meanwhile, automatic analysis and management and control processes are achieved, and the method is suitable for scenes such as classroom teaching management, after-class and extracurricular training, remote examination and family interactive learning.

The invention discloses a multi-modal English learning interaction system and a vocabulary memory training method, and relates to the technical field of English learning, the system comprises the following components: a data acquisition module, a data analysis module, a strategy adjustment module, a resource push module and an interaction learning module; multi-modal learning behavior data, including text input, voice reading, handwritten notes, video learning behaviors, interactive operation and the like, of learners are collected through the data acquisition module, the learners are subjected to group division by applying a group intelligent algorithm, and the behavior pattern and performance of each group in vocabulary learning are analyzed for each group, so that the learning efficiency of the learners is improved. Based on the analysis, the system can automatically adjust teaching strategies and push customized multi-modal learning resources and training methods, so that personalized requirements of different learners are met, and the learning effect and experience are remarkably improved.

The invention discloses a system and device based on interactive learning of multiple doll intelligent agents, and belongs to the technical field of large model intelligent agents, and the system comprises an entity doll which is provided with an NFC label and is used for being placed on a central base and establishing NFC connection with the central base, and the central base carries out NFC induction and recognition on the entity doll and integrates radio and loudspeaker functions. The method comprises the following steps: acquiring and playing an audio signal, performing voice or touch interaction with a user, and establishing communication connection with a service terminal, wherein the service terminal selects a chat theme in combination with a portrait of the user, establishes a virtual chat room, performs role attribute setting on an identified entity doll, identifies the intention of the user, and performs chat interaction with the user; the interactive discussion is output to the central base for playing in combination with the chat theme, so that a plurality of intelligent agents can be supported to deeply discuss around a specific theme, the logic continuity and education consistency of multi-role interactive contents are ensured, real group discussion experience is simulated, and an interactive atmosphere similar to group discussion is formed, thereby improving the interactive quality and learning depth.

The invention relates to the field of industrial automation control, in particular to an intelligent monitoring system oriented to complex working conditions and a reinforcement learning optimization method thereof. Comprising a data acquisition module used for acquiring operation data of industrial equipment under complex working conditions; the data preprocessing module is used for preprocessing the operation data acquired by the data acquisition module and then acquiring standard data; the feature extraction module is used for acquiring key feature vectors capable of reflecting the operation state of the industrial equipment; the intelligent monitoring model building module is used for building an intelligent monitoring model, and the intelligent monitoring model takes the key feature vector as input and outputs a prediction result of the operation state of the industrial equipment after interactive learning; and the alarm module is used for judging whether the industrial equipment is in an abnormal state according to the prediction result, and sending an alarm signal if the industrial equipment is in the abnormal state. According to the method, the problems of low disk monitoring efficiency, poor accuracy, insufficient model adaptability, incomplete alarm mechanism and the like under complex working conditions in the prior art are solved.

The invention provides an imaging department doctor training method and system based on generative artificial intelligence, and relates to the technical field of artificial intelligence, and the method comprises the steps: calling a predetermined case parameterization strategy to carry out the parameterization processing of a real-time case, and obtaining a real-time parameter feature; extracting a first case in the image database, and obtaining a first parameter feature of the first case; generating a virtual case in combination with the real-time parameter features and the first parameter features, and introducing medical knowledge graph analysis to obtain a case reasonable coefficient of the virtual case; if the case reasonable coefficient is in a preset coefficient threshold value, constructing a virtual clinical environment of the virtual case; in a virtual clinical environment, image training is carried out on a target doctor in combination with a virtual case, and the problems that an existing image department doctor training system depends on fixed textbooks and standardized training cases and lacks a dynamic adjustment mechanism based on a real-time case, so that dynamic deduction and immersive interactive learning of doctors are difficult to realize, and the training efficiency is poor are solved. And the technical problem that the image diagnosis skill is difficult to improve efficiently is solved.

The invention discloses an algae abundance dynamic feature selection method based on multi-agent reinforcement learning, which belongs to the technical field of artificial intelligence and data mining, and comprises the steps of multi-source heterogeneous data acquisition and convergence, time sequence data cleaning and standardization processing, driving factor identification and weight calculation key driving factor set establishment, and dynamic feature selection. Convergence and determination of an optimal feature subset selected by multi-agent reinforcement learning and dynamic features: an MARLN system performs multi-round iteration and interactive learning, and each agent continuously optimizes a decision strategy thereof according to a global reward signal fused with a specification item; according to the method, the problems that a traditional static feature selection method cannot adapt to data changes and neglects interaction among features are solved, the most critical driving factors, namely the features, for algae abundance prediction are automatically recognized from multi-source heterogeneous data through a dynamic feature selection method, and the accuracy of algae abundance prediction is improved. And an optimal feature subset is constructed to improve the precision and robustness of the prediction model.

The invention discloses an unmanned aerial vehicle synchronous navigation and radio mapping method based on reinforcement learning and region adaptive integrated learning, and the method comprises the following steps: S1, constructing a radio environment, and carrying out region division according to geographical location information; the method comprises the following steps: S1, detecting a flight area of an unmanned aerial vehicle by adopting an area detector, S4, exploring an environment by the unmanned aerial vehicle through interactive learning, and simultaneously collecting and learning radio frequency data, S5, carrying out strategy training by adopting an experience pool jointly constructed based on actual experience and simulation experience, and S6, storing experience, calculating a return and updating a network. According to the unmanned aerial vehicle synchronous navigation and radio mapping method based on reinforcement learning and regional adaptive ensemble learning, through synchronous navigation and mapping, the unmanned aerial vehicle can be accurately navigated in an unknown or interference environment, meanwhile, a high-precision radio coverage map is established, and the method is particularly suitable for unmanned aerial vehicle communication and navigation tasks in a complex urban environment.

A protein interface prediction method based on three-track coding comprises the following steps: combining a fine-tuned protein language model SiteT5 with evolutionary, geometric and statistical features extracted from a sequence, sending the combined features into a three-track coding network, and integrating a cyclic gating module, a multi-resolution aggregation module and a long sequence deformation module to obtain a protein interface prediction model SiteT5; the method comprises the following steps: respectively capturing a time sequence relation, a local mode and long-range dependence among residues, respectively mapping the three codes into different weights, carrying out point multiplication on the three codes, and carrying out aggregation through a multi-view cross attention module; then the protein residues are sent to a three-layer hierarchical interactive learning module, local structure and global dependency are cooperatively mined through an eight-head gating self-attention module and a position-by-position feedforward module, and finally the probability that each protein residue is an interface is obtained through a classifier. According to the invention, a protein-DNA interface, a protein-RNA interface, a protein-protein interface and an antibody-antigen interface can be effectively captured. And the robustness is ensured, and meanwhile, relatively high prediction precision is also shown.

The invention relates to the field of intelligent temperature control, and provides a refrigeration house PID control parameter optimization method and device, equipment and a storage medium. The method comprises the steps of collecting temperature data and energy consumption data of the refrigeration house at a constant-temperature stage; based on the temperature data and the energy consumption data, a mathematical model representing the mapping relation between the temperature and the energy consumption of the refrigeration house is established; constructing a deep reinforcement learning model based on a preset reward function and a mathematical model giving consideration to the temperature stability and energy consumption optimization of the refrigeration house; and through interactive learning of the intelligent agent of the deep reinforcement learning model and the refrigeration house environment, determining optimization parameters used for adjusting refrigeration house PID control. According to the refrigeration house PID control parameter optimization method, device and equipment and the storage medium provided by the invention, the mathematical model of the temperature and energy consumption mapping relation of the refrigeration house can be effectively established, the deep reinforcement learning model is constructed through the reward function and the mathematical model, and the optimization parameters of refrigeration house PID control are adjusted through deep reinforcement learning.

The invention belongs to the field of multi-room negative pressure gradient control, and discloses a multi-room negative pressure gradient automatic control method and system based on reinforcement learning. An environment model containing multi-dimensional dynamic parameters such as an environment pressure state, an access control state, a fan air volume and a room coupling relation is constructed, so that the method can directly obtain an optimal control strategy through interactive learning on the premise of not needing an accurate physical model. Through the self-adaptive decision-making capability of reinforcement learning, linkage adjustment of the exhaust air volume and the fresh air volume of each room and the power of a total exhaust fan and a fresh air fan of the whole system is achieved, so that the stability of gradients of a corridor-5Pa, a ward-10Pa, a semi-polluted area-15Pa, a polluted area-20Pa and the like is kept, and it is ensured that the pressure difference of all areas is always within the safety range of-7Pa to-2Pa.

The invention belongs to the technical field of auxiliary interactive learning, and particularly relates to a battery-grade lithium carbonate industrial chain monitoring and cognitive deduction system based on industrial intelligence, which comprises a resource management and control database, a knowledge graph modeling library, a situation deduction processing module, a service collaborative application module and a safety protection module, and the resource management and control database is used for storing cross-domain multidisciplinary chain type ubiquitous resource data. According to the battery-grade lithium carbonate industrial chain monitoring and cognitive deduction system based on industrial intelligence, through the resource management and control database and the multi-task distributed scheduling engine, the resource integration and dynamic scheduling capability, the cloud side computing power adaptive scheduling and multi-task parallel processing mechanism are significantly improved; the computing power distribution of the edge end lightweight chip and the cloud high-performance cluster is optimized, the problems of low resource access efficiency and delayed real-time data analysis are reduced, the collaborative management and control of large-scale enterprise resources are realized, and the real-time dynamic monitoring requirement of a complex industrial chain is met.

The invention discloses an equipment residual life prediction method fusing dynamic decomposition and a Koopman operator, and the method comprises the steps: decomposing the preprocessed multi-sensor time sequence data into a trend term and an unstable term through a time sequence decomposition module, and forming a comprehensive trend representation; establishing a Koopman-GRU module, extracting time sequence dependence characteristics in the comprehensive trend representation through the GRU module, and performing linear modeling on the time sequence dependence characteristics in a high-dimensional space by using a Koopman operator module to obtain a space-time state tensor; on the basis of a channel interactive learning module, global core representation between channels is used for capturing the dependency relationship between the sensors, and the feature expression ability is enhanced; generating a final RUL prediction result by using a linear layer of a residual structure in the prediction module; compared with the prior art, the method has the advantages that the feature representation of the multi-sensor time series data is more sufficient, dynamic aggregation and redistribution of the spatial-temporal features are achieved, and then accurate prediction of the nonlinear degradation process of multi-source data monitoring is achieved.

The invention discloses a digital education teaching system and method, and the system comprises a cloud unit, a network transmission unit and an application unit which are in communication connection through a network. An optimization and integration module of the cloud unit fuses voice, image and text multi-modal analysis, accurately locates knowledge point boundaries, realizes personalized output in combination with a'teaching target-editing strategy-student data 'linkage mechanism, and unifies video character styles through an AI technology; the network transmission unit adopts an SHA-256 Hash check algorithm and a Q value comprehensive scoring mechanism to guarantee the transmission quality and dynamically switch communication modes; and the application unit realizes interactive learning and concentration monitoring. According to the invention, the problems of fragmentization of traditional teaching resources, low editing efficiency, insufficient individuation, poor transmission reliability and the like are solved, and the teaching effect and the concentration of students are remarkably improved.

The invention provides a marine image data identification method and system. The method comprises the following steps: step 1, inputting an original marine image; 2, a visual feature extraction module extracts hierarchical features through a multi-scale convolutional neural network; 3, pre-training an LLM semantic reasoning hierarchical attention mechanism semantic mapping multi-task cooperative training framework by the language model agent module; and 4, a dynamic strategy optimization module reinforces learning optimization strategies to adjust identification parameters in real time, and generates a preliminary identification result. And 5, when the user needs to optimize, combining user feedback, a data enhancement interactive learning mechanism and a semantic consistency data enhancement optimization model to realize model updating and parameter adjustment, and returning to the step 2 again. And when the user does not need optimization, outputting a final identification result of the ocean image. According to the invention, a visual coding technology is combined with the semantic understanding capability of LLM, and a recognition solution capable of adaptively processing a multi-modal ocean image is constructed.

The invention relates to the technical field of learning activity optimization, and discloses an interactive learning activity dynamic optimization method and system, and the method comprises the steps: obtaining learning data of a user, extracting an interest label and a learning capability label of the user through a preset user behavior analysis model, and generating a user state label; generating a corresponding activity generation path based on the user state tag in combination with a preset course knowledge base; matching a corresponding learning activity according to the activity generation path to obtain an initial learning activity; optimizing the activity generation path by adding a path branch, trimming the path branch or replacing a path node in combination with a real-time learning state of the user to obtain an activity optimization path; optimizing the initial learning activity through the activity optimization path to obtain an optimized learning activity so as to dynamically optimize the interactive learning activity; according to the method, the learning activity can be dynamically optimized, and the state change condition of the user in the learning process can be responded in real time.

The invention discloses a lightweight multi-agent motion prediction method combined with physical information, and the method comprises the steps: carrying out the vectorization coding of historical tracks and map information of traffic participants, and defining a plurality of local regions according to the position of each agent; introducing physical information to carry out local region feature aggregation on each agent; comprehensively considering road information, and carrying out global interaction on local areas; and outputting a future trajectory of the multiple agents based on a global interaction result. According to the method, the hierarchical design of local interaction modeling and global interaction modeling is used, the calculation redundancy is reduced, the calculation complexity is reduced, and the method is more suitable for large-scale multi-target prediction tasks. During interactive learning, physical priori and physical constraints are introduced, so that wrong learning of invalid relationships is avoided.

The invention discloses a frequency-based self-adaptive interactive learning semi-supervised medical segmentation method and system, and belongs to the technical field of medical image segmentation, and the method comprises the steps: obtaining a to-be-segmented image, inputting the to-be-segmented image into a pre-trained semi-supervised medical segmentation model for processing, and obtaining an image segmentation result; the semi-supervised medical segmentation model comprises a high-frequency sub-network and a low-frequency sub-network, the training process of the semi-supervised medical segmentation model comprises an internal learning process and an external learning process, and the specific training steps are as follows: generating a high-frequency view and a low-frequency view through a frequency double-view stream architecture; in the internal learning process, a high-frequency view and a low-frequency view are used as input to construct a complementary view, and a corresponding sub-network state is updated through cross pseudo tag supervision and mutual uncertainty distance supervision and is used as an initial state of the external learning process; in the external learning process, the corresponding sub-network state is updated through the cross supervision loss and the self-adaptive self-correction loss, and the state is used as the initial state of the internal learning process of the next cycle.

The invention discloses an RAG intelligent knowledge base management method based on a large language model. The method comprises the following steps: analyzing a query context of an enterprise-level user through a situation analysis module, constructing a user portrait, and dynamically adjusting a GraphRAG retrieval and generation strategy; data input by a user is analyzed based on a multi-modal LLM technology, cross-modal retrieval and generation are executed, and knowledge documents or instructions related to multi-modal query are output; through an interactive learning and feedback loop mechanism, revision evaluation of a user on generated content is collected, and revision data is fused to optimize subsequent answer generation; multi-level privacy protection is realized by adopting data desensitization, federal learning and encryption technologies, and answer complexity and professional depth are dynamically adjusted through semantic understanding; the technical problems that multi-modal data processing is difficult, dynamic updating of the knowledge base is insufficient, and retrieval efficiency is low are solved, and meanwhile the functions of business-driven model fine adjustment, knowledge base updating, self-adaptive difficulty grading and privacy protection are achieved.

The invention relates to the field of voice signal processing, in particular to a multi-language interactive learning system based on voice recognition, which is characterized in that sound waves and mouth shape images are synchronously acquired and discretized by the system, and cross-modal coding is formed after alignment; secondly, the code is injected into a micro-ring photon reserve network through phase modulation to unfold time sequence characteristics, the code is mapped into a quaternion graph to be embedded, and segment boundaries are extracted through a diffusion-pulse coupling method; then, according to a fixed field sequence, encapsulating the quaternion graph embedding and segmentation data into an object mark prompt, inputting the object mark prompt into a low-rank adaptive language model, and generating semantic segmentation data and text transcription; and finally, the adaptive learning module implements sparse gradient updating on the low-rank weight and pulse network by using an integer fractal hash exclusive-or difference mask, and adopts support-query element learning for synchronous iteration after user clarification. According to the system, low-power-consumption, high-precision and second-level accent self-adaptive multi-language voice interaction is realized on the end side.

The invention provides an intelligent agent interactive learning and optimization method based on adaptive multi-modal fusion, which relates to the technical field of interactive learning among intelligent agents, comprises links such as multi-modal data preprocessing, adaptive multi-modal fusion, intelligent agent interactive learning and optimization, and realizes multi-modal data processing and intelligent agent performance improvement. Compared with a simple splicing or superficial layer fusion mode of a traditional multi-modal fusion technology, the adaptive multi-modal fusion strategy of the invention can deeply analyze complex internal relations among various modal data such as texts, images, audios and the like. For example, in a scene of combining medical image diagnosis with medical record text analysis, a traditional method is difficult to effectively integrate information of the medical image diagnosis and the medical record text analysis, however, according to the method, association between image features and medical record description is accurately captured by dynamically adjusting a fusion mode, and the accuracy of disease diagnosis is greatly improved. Researches show that after the fusion method is used, the disease misdiagnosis rate is reduced.

The invention provides an early language input and visual training combined interactive learning system, which relates to the field of early education for children and comprises the following modules: an early language input module, an age-based multi-level content system, a visual training module and a tangram and building blocks introduced as visual training tools. A multi-sensory cooperation module is used for encouraging children by using Chinese or English after the children correctly spell corresponding tangram images or build corresponding models, and a Paijiger conservation cognition breakthrough module is used for helping the children to build a length conservation concept and an interactive learning platform before going to school. A personalized learning path is formulated for each child; a feedback and evaluation module; the system is designed according to the cognitive development law of children, early language input and visual training are organically combined, a scientific and systematic training system is formed, and powerful support is provided for future learning and development of children.

The invention discloses a music teaching touch device for blind people, which comprises a supporting seat and a teaching touch seat, and a plurality of dynamic dot matrix seats are arranged on the teaching touch seat in an array. Three rows and two columns of flexible pads are arranged at the top of each dynamic dot matrix base to form a Braille display matrix, and an electromagnetic plunger assembly is arranged below each dynamic dot matrix base to independently drive each dot position to protrude and reset, so that dynamic display of music Braille and graphs is achieved. The bottom of the dynamic dot matrix seat is connected through a silica gel bowl, and a conducting strip is arranged in the bowl and can be in contact with a corresponding conducting material on a circuit board in the teaching touch seat. When a user presses the dynamic dot matrix seat displaying specific notes, pitch sounding of the notes is triggered. According to the invention, the dynamic tactile display function and the pressing input function are integrated, so that refreshing and interactive learning of the braille music score are realized, and a multi-mode feedback closed loop is realized through tactile sense, auditory sense and operation.

The invention relates to the technical field of agent learning, and discloses an agent decision optimization method and device, equipment, a storage medium and a product, and the method comprises the steps: obtaining a to-be-decided task and multi-modal data corresponding to the to-be-decided task; based on the to-be-decided task and a pre-established adaptive fusion model, performing multi-modal fusion on the multi-modal data to obtain multi-modal fusion data; the adaptive fusion model comprises a fusion strategy library formed by a plurality of fusion strategies; inputting the to-be-decided task and the multi-modal fusion data into an intelligent agent for intelligent agent decision making, and obtaining an intelligent agent decision result; wherein the intelligent agent dynamically adjusts a learning strategy and a behavior mode according to environment feedback and interaction with other intelligent agents. According to the invention, through deep mining of the multi-modal data and efficient interactive learning between the agents, the decision-making accuracy of the agents is significantly improved.