首页|Brain-inspired dual-pathway neural network architecture and its generalization analysis

Brain-inspired dual-pathway neural network architecture and its generalization analysis

扫码查看
原文链接
  • 万方数据
  • 维普
In this study,we explored the neural mechanism of global topological perception in the human visual system.We showed strong evidence that the retinotectal pathway in the archicortex of the human brain is responsible for global topological perception,and for modulating the local feature processing in the classical ventral visual pathway.Inspired by this recent cognitive discovery,we developed a novel CogNet architecture to emulate the global-local dichotomy of human visual cognitive mechanisms.The thorough experimental results indicate that the proposed CogNet not only significantly improves image classification accuracies but also effectively addresses the texture bias problem observed in baseline CNN models.We have also conducted mathematical analysis for the generalization gap for general neural networks.Our theoretical derivations suggest that the Hurst parameter,a measure of the curvature of the loss landscape,can closely bind the generalization gap.A larger Hurst parameter corresponds to a better generalization ability.We found that our proposed CogNet achieves a lower test error and attains a larger Hurst parameter,strengthening its superiority over the baseline CNN models further.

global topological perceptiondual-pathwaygeneralization gap analysisHurst parameter

DONG SongLin、TAN ChengLi、ZUO ZhenTao、HE YuHang、GONG YiHong、ZHOU TianGang、LIU JunMin、ZHANG JiangShe

展开 >

College of Artificial Intelligence,Xi'an Jiaotong University,Xi'an 710049,China

School of Mathematics and Statistics,Xi'an Jiaotong University,Xi'an 710049,China

State Key Laboratory of Brain and Cognitive Science,Institute of Biophysics,Chinese Academy of Sciences,Beijing 100101,China

Institute of Artificial Intelligence,Hefei Comprehensive National Science Center,Hefei 230088,China

University of Chinese Academy of Sciences,Beijing 100049,China

展开 >

National Key Research and Development Project of ChinaNational Natural science Foundation of ChinaNational Natural science Foundation of ChinaShenzhen Key Technical ProjectsChinese Academy of SciencesChinese Academy of Sciences

2020AAA0105600U21B204862276208CJGJZD20220517141605012021091YSBR-068

2024

10.1007/s11431-024-2753-3

中国科学:技术科学(英文版)
中国科学院

中国科学:技术科学(英文版)

CSTPCDEI
影响因子：1.056
ISSN：1674-7321
年,卷(期)：2024.67(8)
  • 1