42 results about "Memory retention" patented technology

Systems, devices and methods are provided for stimulating one or more nerves within a user. A system comprises stimulator having an electrode configured for contacting the outer skin surface at a target location and an energy source coupled to the stimulator, The energy source is configured to generate at least one electrical impulse and to transmit the electrical impulse transcutaneously from the electrode through the outer skin surface of the user to a selected nerve in the user. In certain aspects, the systems and methods are particularly useful for non-medical purposes, such as reducing stress and anxiety, enhancing relaxation, improving sleep, energy, concentration and mood, increasing mental or physical performance, promoting mental health, recovery and wellness, generally improving the health and wellbeing of a user and / or for temporarily or permanently improving intelligence, learning capacity, memory retention, recall, mood, alertness and / or sleep efficiency in human beings.

The invention discloses a family-school interaction foreign language teaching system and method based on multi-modal artificial intelligence, and the system comprises the steps: constructing a student personalized interest label library, a curiosity engine library, and a dynamic memory curve and emotion curve model, and fusing the technologies of natural language processing (NLP), computer vision (CV), etc. And a closed-loop teaching system of interest-driven content generation, curiosity-guided learning, memory strategy intelligent adaptation, emotion real-time adjustment and family-school cooperative intervention is formed. Based on a multi-interest fusion generation algorithm of a deep neural network and a generative adversarial network, personalized texts are dynamically constructed, and new grammar and new words are led out through a curiosity engine; experimental data shows that the memory retention rate is increased by 47%, the home-school collaborative foreign language learning efficiency is increased by 2.3 times, the problems of disjunction of interest, low memory efficiency, missing of forgetting curves and emotion management and insufficient home-school collaboration in traditional foreign language teaching are solved, and data privacy security is guaranteed through a Federated Learning framework.

The invention discloses a memory optimization system for a large-scale language model, and the system comprises a dynamic semantic memory modeling module which is used for generating a context-aware memory vector through a bidirectional Transform, and supporting the increment to be updated to one hundred million entity records; the self-adaptive memory resource scheduling module dynamically allocates memory pool partitions based on inference load, high-frequency memory is stored in a working area, and quantitative compression is started in an archiving area; the memory attenuation and enhancement module adopts a two-factor attenuation mechanism to execute enhanced storage and regular pruning on the high-value memory; the cross-scene memory migration module is used for realizing cross-domain adaptation through a domain adapter and comparative learning; and the memory conflict resolution module identifies conflicts based on the rule base and adopts a multi-factor voting mechanism to resolve the conflicts. According to the method, the memory retention integrity and the retrieval efficiency are improved, the resource utilization is optimized, the cross-scene adaptation capability is enhanced, the memory conflict is eliminated, and the model reliability is improved.

The invention relates to the technical field of teaching, and discloses a handwriting recognition method, which comprises the following steps that S1, Chinese character writing specifications are dynamically displayed through AR, a user participates in practice in real time, and the practice is analyzed in real time; s2, a multi-sensory collaborative teaching practice module which synchronously integrates visual, auditory and tactile feedbacks; s3, forming a review strategy based on an intelligent memory enhancement mechanism of the user behavior data; s4, gamifying an interactive interface, and supporting real-time stroke detection and error correction; and S5, a personalized learning path generation module adapted to the learning progress and error mode of the user. According to the method, the writing track of the user is analyzed in real time through the double-branch convolutional neural network, three-level error classification is output, and AR dynamic feedback is combined, so that the stroke sequence detection rate is 92%, the font analysis delay is less than 20 ms, and the error correction efficiency is improved; the multi-sensory collaboration module dynamically distributes visual, auditory and tactile weights by using an attention mechanism to adapt to different learning preferences, so that the memory retention rate is improved, and the environment adaptive capacity is enhanced.

The invention relates to a long text memory keeping method, system and equipment based on a large language model and a storage medium. The method comprises the steps that text contents in long text information are graded through a word segmentation device in a large language model, and all levels of all the text contents are obtained; performing context information weight calculation on the text content of each level based on an attention mechanism to obtain a weight corresponding to each level; storing each text content into a short-term memory space or a long-term memory space according to the weight corresponding to each level; and carrying out compression processing on the short-term memory space or the long-term memory space by using the KV Cache to obtain text content reserved after compression, and outputting the text content. According to the method, the text content in the long text information is graded, and then the weight of the corresponding level is calculated, so that the logic consistency of the context content of multiple dimensions is ensured; the KV Cache is used for carrying out compression processing on the memory space, so that the reasoning time of the large language model can be shortened.

The embodiment of the invention discloses a communication method, device and equipment between Jailhouse virtual machines and a medium, and the method comprises the steps: adding a VNIC module in a Jailhouse source code, and enabling the VNIC module to analyze VNIC configuration and achieve address space mapping, thereby achieving the management of the VNIC; modifying a device tree file of each virtual machine operating system, and adding a memory retention statement which is used for address mapping of the VNIC; driving compatible information is added in the device tree file of each virtual machine operating system, a virtual area mapped by the VNIC is registered based on the memory reservation declaration, and a VNIC interrupt signal is configured; setting a VNIC MAC address in the device tree file according to the number of the virtual machine, and setting a specific bit of the MAC address as the number of the virtual machine for identifying a node to which the communication data flow belongs; and according to the VNIC interrupt signal of the device tree file, modifying an interrupt signal in a memory registration structure body in the configuration file to be consistent with the VNIC interrupt signal of the device tree file, and correspondingly setting an actual memory region mapped by the virtual region.

The invention relates to a dictionary pen learning assistance method and device, equipment and a medium, and the method comprises the steps: extracting learning entries inquired by a user in a target time sequence, and determining the user features and word attribute features of each learning entry; constructing a multi-dimensional feature vector of each learning entry; substituting the multi-dimensional feature vector into an established forgetting curve model, and calculating a corresponding forgetting state; according to the forgetting state, a corresponding word learning prompt strategy is dynamically generated, and the word learning prompt strategy is used for automatically pushing corresponding review prompt information at the target review time point; and obtaining a response behavior of the review prompt information, determining a corresponding state field according to the response behavior, and updating a corresponding word learning prompt strategy according to the state field. By introducing the forgetting curve model to dynamically generate the learning prompt strategy, intelligent review recommendation for the personal memory rhythm of the user can be realized, so that the learning efficiency and the memory keeping effect are improved.

According to the real-time multi-level tile map synthesis method provided by the invention, five-level processing pipelines are not simply parallel, and independent segmentation factor level design, lock-free cross-level data channels and three-level synchronous barrier architecture are carried out; a dynamic memory release mechanism driven by a global tile counting dictionary and a boundary tile divide-conquer synthesis algorithm are used. Through global tile counting dictionary monitoring, the memory is released once the tiles are completed, and accumulation and occupation of the memory are avoided; a boundary adaptive threshold value avoids memory retention caused by the fact that boundary tiles wait to be filled up; five layers of independent treatment pipelines are designed, and the actually measured treatment speed is increased by 4-5 times compared with that of a single-stage serial system; synchronous barriers are arranged in the three stages of initialization, real-time processing and termination, and data consistency is ensured; 1: 1 to 1: 64 times of dynamic scaling is supported through five-level segmentation factors, original features of the image are kept while scaling is carried out by using a nearest neighbor interpolation algorithm, and coordinate range inspection is carried out.

The utility model discloses a mind mapping sounding device for a cardiac pacemaker implanted patient, relates to the technical field of medical auxiliary equipment, and aims to solve the problems that in the prior art, traditional patient education depends on written materials or oral explanation, information is fragmented, the memory retention rate of the patient is low, and postoperative follow-up visit content is complex and is difficult for the patient to remember. Comprising a flexible map guide board and a control box, the upper end of the flexible map guide board is provided with a middle node, a first node and a second node, one side of the first node and one side of the second node are provided with a touch sensing area, the control box is provided with a prompt lamp and a voice interaction module, and the voice interaction module comprises a voice output module and a voice input module. A master control module, a storage module, a data interaction module and a follow-up time management module are further installed in the control box. Operation knowledge, postoperative care, follow-up visit nodes and the like are integrated through mapping, the cognitive level and follow-up visit compliance of a patient are remarkably improved, and the occurrence rate of complications is reduced.

The invention discloses a learning planning method. The method comprises the following steps: acquiring the correct rate of each evaluation dimension of a knowledge point corresponding to an nth review test event; according to the importance of each evaluation dimension, fusing each correct rate to obtain a comprehensive mastering degree P corresponding to the nth review test event; calling a time interval tn; performing memory ability analysis processing based on the comprehensive mastering degree P and the time interval tn to obtain memory intensity S corresponding to the nth review test event; acquiring a preset memory retention rate R; and carrying out time information conversion processing based on a preset memory retention rate R and the memory intensity S corresponding to the nth review test event to obtain a suggested interval duration of the (n + 1) th review test event. According to the method, the problem that the review effect is gradually reduced due to the fact that the current learning plan deviates from the memory condition of the user can be solved, and the requirements of partial emphasized dimensions are also considered while efficient review is carried out on each dimension according to the memory condition of the user.

The present application provides a kind of nutritional composition in the application of the product for improving learning, memory and cognitive ability.The nutritional composition includes animal bifidobacterium lactis subsp.CP-9 and 3'-sialyllactose.Animal bifidobacterium lactis subsp.CP-9 and 3'-sialyllactose can synergistically enhance the memory retention ability of the body, reduce the degree of learning and memory errors, and up-regulate the expression of oligodendrocyte transcription factor 2 coding gene, myelin basic protein coding gene and brain-derived neurotrophic factor coding gene, realize the double regulation of myelin formation and neurotrophic in hippocampus region of brain, synergistically improve the neural plasticity of central nervous system through gut-brain axis pathway, and then double improve the learning, memory and cognitive function of the body from the behavior and molecular level.

The invention provides a method for training the visual clue-dependent positioning reward capability of animals by using an animal behavior training and testing system, and the system realizes that one set of system covers multiple types of experiments through the opening and closing logic of an access door of an adjusting arm, the reward region reward putting rule and the free conversion of visual clues. A visual clue-reward association learning scene is constructed, and animals are guided to establish memory association between clues and reward positions. A three-stage visual clue navigation and memory ability testing method is established, the memory retention ability associated with clue rewards is evaluated in the association stage, the spatial memory ability is independently evaluated in the blank stage, and the regulation and control mechanism of visual clues on animal spatial memory is analyzed in the non-association stage.

The application discloses a memory optimization system for a large language model, comprising: a dynamic semantic memory modeling module, which generates context-aware memory vectors through a bidirectional Transformer and supports incremental updating to 100 million entity records; an adaptive memory resource scheduling module, which dynamically allocates memory pool partitions based on inference load, stores high-frequency memories in the working area, and enables quantization compression in the archive area; a memory decay and enhancement module, which adopts a double-factor decay mechanism, performs enhanced storage on high-value memories, and prunes them regularly; a cross-scenario memory migration module, which realizes cross-domain adaptation through a domain adapter and contrastive learning; and a memory conflict resolution module, which identifies conflicts based on a rule base and resolves contradictions using a multi-factor voting mechanism. The application improves memory retention integrity and retrieval efficiency, optimizes resource utilization, enhances cross-scenario adaptation capability, resolves memory conflicts, and improves model reliability.

The application provides a large model knowledge dynamic iteration updating method and system, relates to the technical field of large models, collects multiple types of dynamic knowledge sources in the field of children's tumors and is attached with relevant information, and constructs a knowledge evolution life cycle network based on this; the knowledge content nodes are subjected to multi-source credibility cross verification and dynamic conflict reconciliation processing to obtain a target knowledge set; a parameter fine tuning and memory reservation collaborative mechanism is called to convert into a knowledge updating parameter package; the knowledge updating parameter package is input into a large model to perform collaborative updating operation and generate a knowledge updating whole-process traceability document. The application can comprehensively and accurately process multiple types of knowledge sources, effectively improve the knowledge accuracy and reliability, and realize model collaborative updating.

The invention discloses a holographic foreign language learning intelligent synergy system, and relates to the intelligent education hardware and neural engineering cross technology field, and the system comprises a head wearable communication device, an integrated visual display unit, an auditory feedback unit, a motion capture unit and a central processor; the cloud scene memory strengthening module is used for storing a multi-language dictionary, a grammar database, a scene simulation scene and a real-time interaction program; and the multi-sensory nerve collaborative activation module synchronously stimulates the nerve center of the user through the following paths, and the system has the effect of optimizing memory maintenance: in combination with a scenario review strategy of a memory rule, learning content retention can be reinforced at key nodes, and the frequency of repeated learning is reduced; the virtual scene training provides language interaction experience similar to reality, and the adaptation process from learning to practical application can be shortened.

The application belongs to the field of intelligent teaching, and particularly relates to an English oral pronunciation correction system based on artificial intelligence, which comprises a multi-modal perception module, a deep pronunciation feature extraction module, a cognitive compensation fusion module and an intelligent correction and training planning module; the abstract pronunciation is converted into a tactile experience through a multi-modal learning process, the tactile information and the auditory information are synchronously input, the memory retention rate of the visually impaired students is improved, and the students are guided to deeply remember the acoustic characteristics of the pronunciation; the application simulates the auditory-tactile joint representation ability of the visually impaired brain through the learnable attention weight and the cognitive compensation prior mask, realizes the real sense of "touch instead of vision", simultaneously identifies the error reasons through multi-task learning, provides scientific basis for targeted training, designs a special tactile generation network, maps the acoustic characteristics and the pronunciation error types into tactile patterns, provides instant tactile correction feedback for the students, and forms a closed-loop multi-modal learning experience.

To provide a semiconductor device with a novel structure, in which a memory content can be held even in a situation with no power supply in a memory holding period, and the number of times of writing is not limited.SOLUTION: A semiconductor device includes a first transistor including a first channel formation region including a semiconductor material other than an oxide semiconductor, a second transistor including a second channel formation region including an oxide semiconductor material over the first transistor, and a capacitor element. One of a second source electrode and a second drain electrode of the second transistor is connected electrically to one electrode of the capacitor element.SELECTED DRAWING: Figure 1

The invention discloses a learning planning method, and the method comprises the steps: obtaining a time interval t between the last two learning events, and a test correct rate A corresponding to the last learning event; performing memory ability analysis processing based on the test accuracy A and the time interval t to obtain current memory intensity S; acquiring a preset memory retention rate R; performing time information conversion processing based on a preset memory retention rate R and the current memory intensity S to obtain a suggested interval duration of the next learning event; and according to the end time corresponding to the last learning event and the suggested interval duration of the next learning event, planning the starting time of the next learning event. According to the scheme, the updated memory intensity S is obtained by using the test accuracy A, and based on the updated memory intensity S and the memory retention rate R expected by the user, the learning time planning conforming to the current memory ability and the memory effect expected value of the user is realized, so that the review of the user is more efficient.

A processing system for a lighting arrangement configured to output light when powered by an input power from an external power source. The processing system is configured to control the output of light by the lighting arrangement when also powered by the input power. Particularly, when the processing system is operating in a sequence control mode, the processing system is configured to control the lighting arrangement to output light in accordance with a time-dependent lighting sequence, which describes desired characteristics of the light output by the lighting arrangement over a duration of time.

The invention provides an English vocabulary cognition training method and system based on an interval repetition memory method, and relates to the technical field of language cognition training, abnormal value elimination and normalization preprocessing are performed on collected data, a memory intensity coefficient is calculated, a memory intensity level is judged, and a final review interval is generated in combination with a vocabulary basic interval and a special adjustment rule; each layer is matched with a differentiated multi-modal training task, and the vocabulary adding proportion, cognitive layer dynamic switching and fatigue relieving links in the training process are controlled. According to the method, multi-dimensional data such as learning duration, memory performance and vocabulary difficulty and dynamic review intervals are fused, compared with fixed period review, the one-week memory retention rate of vocabularies is increased to 85%, the two-week memory retention rate reaches 72%, the long-term memory effect is remarkably enhanced, and the problem that multi-dimensional data such as intervals, non-fused reaction time and cognitive states are adjusted only based on correct rate is solved. And personalized and efficient vocabulary cognition training requirements cannot be met.

The application relates to a device monitoring method, device and computer equipment for a substation debugging process. The method comprises the following steps: establishing a source occupation account page for a subscription object, determining a test source takeover or real source recovery event based on real-time message flow, and updating a current occupation source state; when a test stage end signal is received, sending an occupation release probe message to the subscription object occupied by the test source; determining an occupation release result according to whether a real source recovery message is listened to and whether the current occupation source state has been overwritten as real source occupation; classifying and mapping subsequent business messages to determine semantic attribution, generating an explanation result and overwriting the account page; based on the occupation release result and the latest message explanation classification, when a recovery reconstruction mark needs to be added, a real source recovery reconstruction section is constructed and inserted into the original next round of test process. The method can identify and eliminate semantic dislocation caused by source memory retention of the subscription object in the smart substation debugging process, and improve the debugging reliability.

The invention relates to the technical field of education, and discloses a multi-sensory linkage educational digital oil painting system which comprises a knowledge visualization sketch layer, an intelligent auxiliary terminal layer and a multi-sensory linkage assembly layer. The knowledge visualization sketch layer is used for presenting a digital oil painting sketch fused with educational information, and a color theory mark, a stroke skill prompt and a famous painting background knowledge pop-up window are embedded in the sketch; and the intelligent auxiliary terminal layer comprises a touch operation module, a data processing module and an educational resource library. In the invention, the knowledge visualization sketch layer is embedded with education information such as a color theory and stroke skills through dynamic labeling, audio explanation, tactile feedback and scene smell of the multi-sensory linkage component layer are combined, a hippocampus memory area is activated by relying on a right brain image memory advantage and an alpha wave regulation and control technology, and multi-sensory synchronous input is combined, so that a hippocampus memory area is activated. A knowledge conversion path conforming to the Ebbinghaus multi-channel memory model is constructed, and the memory retention rate is remarkably improved.

A distillation enhancement method for class incremental learning comprises the following steps: (1) constructing a class incremental learning baseline model based on sample playback and knowledge distillation, and performing joint training by using current task data and historical memory library samples in each incremental stage; (2) a semantic confusion matching SCM mechanism is introduced, prediction deviation of a teacher model is quantified by constructing a category confusion matrix, soft labels are corrected, and interference of new categories with similar semantics is inhibited; (3) introducing a long-term memory attention distillation (LTMAD) mechanism, and introducing a distillation weight of time distribution index growth according to the category to strengthen the protection of the early category; (4) combining the SCM loss and the LTMAD loss with the original classification loss and the distillation loss to form a multi-objective optimization function; and (5) carrying out model training and evaluation by using a CIFAR100 type incremental learning data set and an Image Net-Subset type incremental learning data set. According to the invention, the memory retention capability of the class incremental learning model is improved.

The invention provides an intelligent content pushing and interaction method and device, and relates to the field of computers. The method comprises the following steps: constructing a target user portrait based on multi-dimensional user data; based on the target user portrait and the historical learning data, predicting a forgetting time point of each learning unit, and obtaining predicted memory data; wherein in the prediction process, an Ebbinghaus attenuation correction factor is introduced to dynamically adjust predicted data; the Ebbinghaus attenuation correction factor is generated based on the deviation between the actual memory retention rate of the consolidation test of the target user and the theoretical memory forgetting curve; based on the predicted memory data, a grey wolf algorithm is utilized to dynamically optimize multiple combined learning strategies, and an optimal combined learning strategy is determined; and based on the optimal combined learning strategy, establishing a pushing rule, and based on the pushing rule, pushing the corresponding learning content to a user terminal of the target user. According to the invention, intelligent content pushing can be realized, and the real-time performance and accuracy of pushing and the overall performance of the system are improved.

The invention provides an unsupervised cross-modal retrieval method based on memory preserving Hash, and relates to the technical field of unsupervised cross-modal retrieval, comprising the following steps: feature extraction and Hash learning; feature fusion and similar matrix construction; and training and optimizing the model. According to the method, the historical memory similar matrix construction module is used for providing stable training supervision, so that the discrimination capability and retrieval robustness of the Hash codes are effectively improved. According to the method, semantic tags are not needed, the method is suitable for unmarked multimedia data retrieval in a real scene, the model structure is light, the retrieval efficiency is high, and the method has high practical application capacity and expansibility.

Systems, methods and devices are provided for delivering a combination of electrical impulses (and / or fields) and various stimuli to bodily tissues for various purposes. The combined therapy increases the brain's capacity to clear metabolic and neurotoxic material (waste removal), facilitates neuroplasticity to affect brain network optimization (learning and memory consolidation), and restores brain energy and neurotransmitter levels (neurometabolic restoration). In certain aspects, the systems and methods are particularly useful for enhancing sleep quantity, quality, and / or efficiency to promote more effective sleep, including more efficient waste clearance, memory consolidation, neurotransmitter rebalancing, and maintenance of energy homeostasis, leading to enhanced neurological health and reduced required sleep duration. The systems and methods may also be useful for improving a cognitive performance of the user by maintaining cognitive function during sleep deprivation and / or temporarily or permanently improving intelligence, learning capacity, memory retention, recall, mood and / or alertness.

The invention discloses an intelligent association review method, a storage medium and equipment, and the method comprises the following steps: S1, constructing a knowledge graph network which comprises a plurality of knowledge point nodes and edges representing the relationship between knowledge points; s2, clustering a plurality of knowledge points to form a knowledge cluster according to a relationship among the knowledge points in the knowledge graph network; s3, calculating a comprehensive forgetting degree of each knowledge cluster, wherein the comprehensive forgetting degree is determined based on a memory retention value and an association weight of each knowledge point in the knowledge cluster; s4, generating an optimal review path according to the comprehensive forgetting degree of each knowledge cluster, wherein the review path comprises a review sequence of the knowledge clusters and a combination mode of associated knowledge points; and S5, dynamically adjusting the review interval of the knowledge cluster according to the actual review effect of the user. According to the method, more scientific memory consolidation can be realized by constructing the knowledge point network, calculating the knowledge cluster forgetting degree and dynamically optimizing the review path.

Certain aspects of the present disclosure are directed towards techniques and apparatus for wireless communication. An example method generally includes: storing one or more communication settings for signal communication in memory; performing a first signal communication using the one or more communication settings; entering an idle mode of operation after performing the signal communication; retaining the one or more communication settings in the memory during the idle mode; and performing a second signal communication after the idle mode using the one or more communication settings retained in the memory.

A system for assisting memory retention, consolidation and recall. The system has a first apparatus / subsystem for acquiring first information, and a first process that executes on the first apparatus / subsystem or a connected second apparatus / subsystem to use the first information to form second information associated with the first information. A second process executes on the first apparatus / subsystem or a connected second apparatus / subsystem to use the second and first information to create associations between parts of the first information and store these associations with the first information so that this information can be updated when new first information is made available. The first and third information is made relational to time, and a fourth information, which is formed according to a first algorithm and made relational to time according to a second algorithm, is stored in a way that allows it to be updated when new first and second information becomes available. The fourth information is presented as output as per time relation created by the second algorithm via the first, second, or a connected third apparatus / subsystem to a user, and repeated at a time relation set by a third algorithm that uses the first information.

The envisioned Artificial Intelligence Long-term Memory System presents a significant advancement in artificial intelligence (AI) by addressing the challenge of transient memory in AI language models. This innovation introduces a hardware-centric approach to augment the memory faculties of AI language models, enabling them to store, access, refine, and incorporate specific data points for deeper user engagement. The enhancements focus on long-term memory, facilitating AI models to remember and build upon past interactions, thus offering a more natural and intuitive interaction between computers and users. The potential applications span from personal computing to complex medical diagnostics, marking a pivotal step towards AI models functioning with contextual awareness and memory retention akin to human interaction.