Kirsch, L; Kunze, J; Barber, D; (2018) Modular Networks: Learning to Decompose Neural Computation. In: Bengio, S and Wallach, H and Larochelle, H and Grauman, K and Cesa-Bianchi, N and Garnett, R, (eds.) Advances in Neural Information Processing Systems 31 (NIPS 2018). NIPS Proceedings: Montreal, Canada.

Preview

Text Kunze_7508-modular-networks-learning-to-decompose-neural-computation.pdf - Published Version Download (1MB) | Preview

Abstract

Scaling model capacity has been vital in the success of deep learning. For a typical network, necessary compute resources and training time grow dramatically with model size. Conditional computation is a promising way to increase the number of parameters with a relatively small increase in resources. We propose a training algorithm that flexibly chooses neural modules based on the data to be processed. Both the decomposition and modules are learned end-to-end. In contrast to existing approaches, training does not rely on regularization to enforce diversity in module use. We apply modular networks both to image recognition and language modeling tasks, where we achieve superior performance compared to several baselines. Introspection reveals that modules specialize in interpretable contexts.

Download activity - last month

Export as CSVJSONXML

Download activity - last 12 months

Export as CSVJSONXML

Downloads by country - last 12 months

Export as XMLJSONCSV

Archive Staff Only

View Item