Towards biologically plausible DNN optimization: Replacing backpropagation and loss functions with a top-down credit assignment network
Authors:
- Jianhui ChenJCJianhui Chen
- Tianming YangTYTianming Yang
- Cheng-Lin LiuCLCheng-Lin Liu
- Zuoren WangZWZuoren Wang
- 5 December 2025
- journal article
- Published by Elsevier in Neurocomputing
- Vol. 667
Abstract
No abstract availableFunding Information
- Chinese Academy of Sciences (XDB1010201)
- Ministry of Science and Technology of the People's Republic of China (2022ZD0205100)
This publication has 37 references indexed in Scilit:
- Smooth momentum: improving lipschitzness in gradient descentApplied Intelligence, 2022
- Top-down control of hippocampal signal-to-noise by prefrontal long-range inhibitionCell, 2022
- My Neighbour Hetero — deconstructing the mechanisms underlying heterosynaptic plasticityCurrent Opinion in Neurobiology, 2021
- 2019 Evolutionary Algorithms Review2020
- Mechanisms and regulation of dopamine releaseCurrent Opinion in Neurobiology, 2019
- Dendritic solutions to the credit assignment problemCurrent Opinion in Neurobiology, 2019
- Top-Down Neural Attention by Excitation BackpropInternational Journal of Computer Vision, 2017
- The role of 5-HT1A receptors in learning and memoryBehavioural Brain Research, 2008
- Diffusional extrasynaptic neurotransmission via glutamate and GABANeuroscience and Behavioral Physiology, 2005
- Multilayer feedforward networks are universal approximatorsNeural Networks, 1989