Mitigating Deep Double Descent by Concatenating Inputs

John Chen, Qihan Wang, Anastasios Kyrillidis

Keywords: deep double descent, feedforward neural network, image classificaiton

Abstract: The double descent curve is one of the most intriguing properties of deep neural networks. It contrasts the classical bias-variance curve with the behavior of modern neural networks, occurring where the number of samples nears the number of parameters. In this work, we explore the connection between the double descent phenomena and the number of samples in the deep neural network setting. In particular, we propose a construction which augments the existing dataset by artificially increasing the number of samples. This construction empirically mitigates the double descent curve in this setting. We reproduce existing work on deep double descent, and observe a smooth descent into the overparameterized region for our construction. This occurs both with respect to the model size, and with respect to the number epochs.

One-sentence Summary: We introduce a construction to artificially increase the size of the dataset which mitigates deep double descent in a variety of settings.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Reviewed Version (pdf): https://openreview.net/references/pdf?id=siCL0eYoztL