Aligning Artificial Neural Networks to the Brain yields Shallow Recurrent Architectures

Jonas Kubilius, Martin Schrimpf, Ha Hong, Najib J. Majaj, Rishi Rajalingham, Elias B. Issa, Kohitij Kar, Pouya Bashivan, Jonathan Prescott-Roy, Kailyn Schmidt, Aran Nayebi, Daniel Bear, Daniel L. K. Yamins, James J. DiCarlo

Sep 27, 2018 ICLR 2019 Conference Blind Submission readers: everyone Show Bibtex
  • Abstract: Deep artificial neural networks with spatially repeated processing (a.k.a., deep convolutional ANNs) have been established as the best class of candidate models of visual processing in the primate ventral visual processing stream. Over the past five years, these ANNs have evolved from a simple feedforward eight-layer architecture in AlexNet to extremely deep and branching NASNet architectures, demonstrating increasingly better object categorization performance. Here we ask, as ANNs have continued to evolve in performance, are they also strong candidate models for the brain? To answer this question, we developed Brain-Score, a composite of neural and behavioral benchmarks for determining how brain-like a model is, together with an online platform where models can receive a Brain-Score and compare against other models. Despite high scores, typical deep models from the machine learning community are often hard to map onto the brain's anatomy due to their vast number of layers and missing biologically-important connections, such as recurrence. To further map onto anatomy and validate our approach, we built CORnet-S: an ANN guided by Brain-Score with the anatomical constraints of compactness and recurrence. Although a shallow model with four anatomically mapped areas and recurrent connectivity, CORnet-S is a top model on Brain-Score and outperforms similarly compact models on ImageNet. Analyzing CORnet-S circuitry variants revealed recurrence as the main predictive factor of both Brain-Score and ImageNet top-1 performance.
  • Keywords: Computational Neuroscience, Brain-Inspired, Neural Networks, Simplified Models, Recurrent Neural Networks, Computer Vision
0 Replies

Loading