Abstract
AbstractMachine learning (ML) models have long overlooked innateness: how strong pressures for survival lead to the encoding of complex behaviors in the nascent wiring of a brain. Here, we derive a neurodevelopmental encoding of artificial neural networks that considers the weight matrix of a neural network to be emergent from well-studied rules of neuronal compatibility. Rather than updating the network’s weights directly, we improve task fitness by updating the neurons’ wiring rules, thereby mirroring evolutionary selection on brain development. We find that our model (1) provides sufficient representational power for high accuracy on ML benchmarks while also compressing parameter count, and (2) can act as a regularizer, selecting simple circuits that provide stable and adaptive performance on metalearning tasks. In summary, by introducing neurodevelopmental considerations into ML frameworks, we not only model the emergence of innate behaviors, but also define a discovery process for structures that promote complex computations.
Funder
U.S. Department of Health & Human Services | NIH | National Institute of General Medical Sciences
Publisher
Springer Science and Business Media LLC
Subject
General Physics and Astronomy,General Biochemistry, Genetics and Molecular Biology,General Chemistry,Multidisciplinary
Reference55 articles.
1. Musall, S., Urai, A. E., Sussillo, D. & Churchland, A. K. Harnessing behavioral diversity to understand neural computations for cognition. Curr. Opin. Neurobiol. 58, 229 (2019).
2. Richards, B. A. et al. A deep learning framework for neuroscience. Nat. Neurosci. 22, 1761 (2019).
3. Srinivasan, S., Greenspan, R. J., Stevens, C. F. & Grover, D. Deep (er) learning. J. Neurosci. 38, 7365 (2018).
4. Zador, A. M. A critique of pure learning and what artificial neural networks can learn from animal brains. Nat. Commun. 10, 1 (2019).
5. Lillicrap, T. P. & Kording, K. P. What does it mean to understand a neural network? arXiv Prepr. arXiv 1907, 06374 (2019).
Cited by
9 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献