4 results about "Memory acquisition" patented technology

This invention belongs to the field of pharmaceutical technology, specifically relating to the application of a traditional Chinese medicine composition in the preparation of drugs for improving cognitive impairment. By weight, the traditional Chinese medicine composition comprises: 225-300 parts rhubarb, 50-75 parts pig bile powder, 90-120 parts aster, 360-480 parts scrophularia, 180-240 parts immature bitter orange peel, and 180-240 parts areca nut. This invention is the first to demonstrate that a composition of six traditional Chinese medicines, including scrophularia, rhubarb, and areca nut, significantly improves memory acquisition (induced by scopolamine), consolidation (induced by sodium nitrite), and retrieval impairments, covering the entire process of memory formation, encoding, storage, and retrieval, thus overcoming the limitations of existing drugs that only target a single memory stage.

The invention discloses a self-adaptive long-term memory management system and method for end-side AI hardware, and belongs to the technical field of artificial intelligence and edge computing. The system comprises a memory acquisition module, a memory coding module, a memory storage module, a memory index module, a self-adaptive retrieval module, a memory evaluation module, a self-adaptive elimination module, a resource monitoring module, a strategy scheduling module and a privacy protection module. Through a hierarchical storage architecture, a multi-dimensional index structure and a mixed retrieval strategy, large-scale long-term memory is efficiently managed on end-side equipment with limited resources; dynamically adjusting a management strategy according to an equipment resource state and a user behavior mode through a self-adaptive elimination mechanism and strategy scheduling; and the privacy security of the user is guaranteed through encrypted storage and access control. According to the method, the memory management efficiency and the intelligent level of the end-side AI equipment are remarkably improved, the storage efficiency is improved by 40% or above, the retrieval accuracy is improved by 30%, the retrieval delay is reduced by 50%, the power consumption is reduced by 60%, and the method is suitable for various end-side AI application scenes such as smart phones, IoT (Internet of Things) equipment and vehicle-mounted systems.

The application relates to the technical field of artificial intelligence, and discloses an intelligent reply method and device, equipment and a medium, which belong to the technical field of artificial intelligence, are applied to financial and medical technology scenes, and comprise the following steps: acquiring a hierarchical business memory library containing short-term, medium-term and long-term business memory; searching candidate business memory from the business memory library according to a business demand category; acquiring a candidate memory time and performing time-effect weight calculation with a current time; screening target business memory according to the calculated memory time-effect weight; constructing business question indication information by combining the target business memory with user business question data, so as to instruct a business intelligent agent to generate reply data. In the application, candidate memory is preliminarily searched from the hierarchical business memory library according to business demand, and target memory is screened in combination with time-effect weight, so that reply data is generated based on the target memory, the reply search accuracy is improved, and the reply content generated by the business intelligent agent is more time-effective and reliable.

The invention discloses a method and system for generating granules with regular internal structures and completely tangent boundaries, and the method comprises the steps: obtaining a function through a memory according to the shape of an internal structure of a granule model to be constructed, and constructing an initial closed front-end chain; based on the particle pairs in the scanning closed front-end chain, determining to-be-generated particles which are tangent to the particle pairs and have the particle size of R0 according to a geometric principle; performing overlapping judgment and particle overlapping processing on the to-be-generated particles and the generated particles until the to-be-generated particles and the current particles have no overlapping phenomenon; and performing boundary approximation judgment and boundary tangency processing on the to-be-generated particles and the model boundary, and generating final new particles by adopting a particle generation command. And when the radiuses of all the to-be-generated particles are smaller than the termination critical radius, a final particle numerical model is obtained. According to the method, the problem that particles with regular-shaped internal structures are attached to boundaries is solved, and efficient generation of the particles and effective application of boundary conditions are ensured.