The research team of the School of Psychology of the Capital Normal University published research results in the top international journal "Advanced Science"

3 min read
Professor Liang Peng of the School of Psychological University of the Capital Normal University and Professor Zhong Ning (Foreign Academician of the Japanese Academy of Engineering, special professor of the Capital Normal University) team, Beijing University of Technology, Australia South Queensland University in the International Academic Journal “Advanced Science”Advanced Science, Q1, If =” 15.1) published a research paper entitled “NEVER-EENDING” Learning for Explainable Brain Computing. “This study proposes an explanatory multi-source brain big data intelligent computing framework. Based on human participation, the NEVER-ending Learning paradigm, integrate the combination of evidence-fusion calculation method to knowledge-information-data-INFORMATION-DATA, Kid) Study and verify that it is applied to inference brain research.
The study combines the internal evidence learning of multi -tasking functional image data analysis, and the external evidence learning of the analysis of functional imaging literature published publicly, which aims to support the intelligent computing of brain big data that can be explained, generalized, and robust.The core is to jointly use knowledge-driven positive reasoning and data-driven reverse reasoning to never stop learning, and integrate pre-trained language modeling technology and human-machine interaction learning (Human-in-The-loop) mechanism to achieveSystematic analysis of human brain high -level cognitive functions.Specifically, the method framework first learns multi -dimensional design parameters of different cognitive experiments, as well as logical rules and mapping weights between each experiment.Then, integrate multi -source brain data resources according to the rules, and combine the mapping weight calculation of the search resources to obtain uncertain parameter values τ and support parameter value, and then infer the brain structure and cognitive functionsMulti -mapping relationships.Experiments show that the method framework of this method uses the integration calculation and uncertain reasoning of multi -source multi -tasking functional image data, which will help further deepen the understanding of high -level cognitive functions such as inductive reasoning.
The results of this paper help to deeply decoding the complicated information processing mechanism of the human brain, and opened up a new path for systemic computing and multi -dimensional analysis for brain cognitive neuroscience research.Over time, you can use the newly collected internal and external versatile evidence, and never stop learning the paradigm in combination, to achieve the combination of multi-perspective persistenceReasoning is the core of KID (knowledge-information-data) recycling. It is expected to provide new technologies for brain big data intelligent computing technologies for further revealing the inherent characteristics and mechanisms of human intelligence.
This article is the latest research results in Liang Peipeng’s research group and Academician Zhong Ning’s team around the new cross -disciplinary series of brain informatics. Other preliminary results were published in journals and conferences such as “Information Fusion” (IF = 17.564).Essence
Thesis link: https://doi.org/10.1002/ADVS.202307647
(Source: School of Psychology; Author: Liang Peipeng; Editor on duty: Wang Yijie, Yang Xiaofei, Wang Dan, Mei Lanzi)
Report/feedback

You May Also Like

More From Author