Multi-Concurrent Real-time Adversarial System for Reinforcement Learning Training and Evaluation

Abstract

The invention belongs to the technical field of artificial intelligence, and specifically relates to a multi-concurrent real-time confrontation system oriented to reinforcement learning training and evaluation, aiming at solving the problem that the existing confrontation system does not use the memory training mode, so the system is not suitable for the training and training of reinforcement learning methods. Evaluation, so as to combat the problem that the effect of decision-making is not as expected. The invention includes: a confrontation scheduling management module, which creates confrontation sites, confrontation processes, and confrontation scenario parameters according to confrontation requirements; an engine kernel module, combined with deduction personnel or AI action sets, updates deduction status and situation, and generates real-time deduction situation data; deduction users On the terminal, the situational data of real-time deduction is parsed into graphics presented in map grids and displayed, and the operation instructions of the deduction personnel or AI are obtained and action sets are generated; on the counter-observation side, the situational data of real-time deduction is parsed into 3D models and graphics and displayed And switch the display at the set viewing angle. The countermeasure system of the invention has good countermeasure decision-making effect and wide application.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence, and in particular relates to a multi-concurrent real-time confrontation system oriented to reinforcement learning training and evaluation. Background technique [0002] With the development of artificial intelligence technology represented by deep learning, human beings have made great progress in tasks of "perceived intelligence" such as image processing, speech recognition, and text processing. However, "perceptual intelligence" is the ability of machines to obtain information through various sensors. The main defect is that each algorithm is only applicable to specific problems and does not have the complete cognitive ability of humans. In contrast, "cognitive intelligence" refers to the ability of machines to actively think, understand, and reason, and to achieve self-learning, purposeful reasoning, and interaction with the environment without prior programming by humans. Alt...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the above-mentioned problems in the prior art, that is, the existing confrontation system does not use the memory training mode, so the system is not suitable for the training and evaluation of the reinforcement learning method, so that the confrontation decision-making effect cannot reach the expected problem, the present invention provides a A multi-concurrent real-time confrontation system oriented to reinforcement learning training and evaluation, the real-time confrontation system includes an engine kernel module, a confrontation scheduling management module, a deduction client and a confrontation observation terminal;

Images