TY - JOUR
T1 - Explainable neural networks that simulate reasoning
AU - Blazek, Paul J.
AU - Lin, Milo M.
N1 - Funding Information:
We acknowledge the Cecil H. and Ida Green Foundation, the Welch Foundation (grant no. I-1958-20180324) and the anonymous-donor-supported UTSW High Risk/High Impact grant for funding this research.
Publisher Copyright:
© 2021, The Author(s), under exclusive licence to Springer Nature America, Inc.
PY - 2021/9
Y1 - 2021/9
N2 - The success of deep neural networks suggests that cognition may emerge from indecipherable patterns of distributed neural activity. Yet these networks are pattern-matching black boxes that cannot simulate higher cognitive functions and lack numerous neurobiological features. Accordingly, they are currently insufficient computational models for understanding neural information processing. Here, we show how neural circuits can directly encode cognitive processes via simple neurobiological principles. To illustrate, we implemented this model in a non-gradient-based machine learning algorithm to train deep neural networks called essence neural networks (ENNs). Neural information processing in ENNs is intrinsically explainable, even on benchmark computer vision tasks. ENNs can also simulate higher cognitive functions such as deliberation, symbolic reasoning and out-of-distribution generalization. ENNs display network properties associated with the brain, such as modularity, distributed and localist firing, and adversarial robustness. ENNs establish a broad computational framework to decipher the neural basis of cognition and pursue artificial general intelligence.
AB - The success of deep neural networks suggests that cognition may emerge from indecipherable patterns of distributed neural activity. Yet these networks are pattern-matching black boxes that cannot simulate higher cognitive functions and lack numerous neurobiological features. Accordingly, they are currently insufficient computational models for understanding neural information processing. Here, we show how neural circuits can directly encode cognitive processes via simple neurobiological principles. To illustrate, we implemented this model in a non-gradient-based machine learning algorithm to train deep neural networks called essence neural networks (ENNs). Neural information processing in ENNs is intrinsically explainable, even on benchmark computer vision tasks. ENNs can also simulate higher cognitive functions such as deliberation, symbolic reasoning and out-of-distribution generalization. ENNs display network properties associated with the brain, such as modularity, distributed and localist firing, and adversarial robustness. ENNs establish a broad computational framework to decipher the neural basis of cognition and pursue artificial general intelligence.
UR - http://www.scopus.com/inward/record.url?scp=85123712277&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85123712277&partnerID=8YFLogxK
U2 - 10.1038/s43588-021-00132-w
DO - 10.1038/s43588-021-00132-w
M3 - Article
C2 - 38217134
AN - SCOPUS:85123712277
SN - 2662-8457
VL - 1
SP - 607
EP - 618
JO - Nature Computational Science
JF - Nature Computational Science
IS - 9
ER -