Title: Efficient Model Performance Estimation via Feature Histories

Abstract:

An important step in the task of neural network design, such as hyper-parameter optimization (HPO) or neural architecture search (NAS), is the evaluation of a candidate model’s performance. Given fixed computational resources, one can either invest more time training each model to obtain more accurate estimates of final performance, or spend more time exploring a greater variety of models in the configuration space. In this work, we aim to optimize this exploration-exploitation trade-off in the context of HPO and NAS for image classification by accurately approximating a model’s maximal performance early in the training process. In contrast to recent accelerated NAS methods customized for certain search spaces, e.g., requiring the search space to be differentiable, our method is flexible and imposes almost no constraints on the search space. Our method uses the evolution history of features of a network during the early stages of training to build a proxy classifier that matches the peak performance of the network under consideration. We show that our method can be combined with multiple search algorithms to find better solutions to a wide range of tasks in HPO and NAS.

Committee:

Kris M. Kitani (Advisor)

Deva Ramanan

Xiaofang Wang