Presented By
O’Reilly + Intel AI
Put AI to Work
April 15-18, 2019
New York, NY

Random Search and Reproducibility for Neural Architecture Search

Ameet Talwalkar (Carnegie Mellon University and Determined AI)
1:00pm1:40pm Wednesday, April 17, 2019
Machine Learning, Models and Methods
Location: Grand Ballroom West
Secondary topics:  Automation in machine learning and AI, Deep Learning and Machine Learning tools, Models and Methods

Who is this presentation for?

Machine Learning Engineers, Data Scientists

Level

Intermediate

Prerequisite knowledge

Attendees should be familiar with the basics of Machine Learning and Deep Learning.

What you'll learn

We propose new NAS baselines that build off the following observations: (i) NAS is a specialized hyperparameter optimization problem; and (ii) random search is a competitive baseline for hyperparameter optimization. Leveraging these observations, we evaluate both random search with early-stopping and a novel random search with weight-sharing algorithm on two standard NAS benchmarks---PTB and CIFAR-10.

Description

Neural architecture search (NAS) is a promising research direction that has the potential to replace expert-designed networks with learned, task-specific architectures. In this talk, we present work which aims to help ground the empirical results in this field. We propose new NAS baselines that build off the following observations: (i) NAS is a specialized hyperparameter optimization problem; and (ii) random search is a competitive baseline for hyperparameter optimization. Leveraging these observations, we evaluate both random search with early-stopping and a novel random search with weight-sharing algorithm on two standard NAS benchmarks—-PTB and CIFAR-10. Our results show that random search with early-stopping is a competitive NAS baseline, e.g., it performs at least as well as ENAS, a leading NAS method, on both benchmarks. Additionally, random search with weight-sharing outperforms random search with early-stopping, achieving a state-of-the-art NAS result on PTB and a highly competitive result on CIFAR-10. Finally, we explore the existing reproducibility issues of published NAS results. We note the lack of source material needed to exactly reproduce these results, and further discuss the robustness of published results given the various sources of variability in NAS experimental setups. Relatedly, we provide all information (code, random seeds, documentation) needed to exactly reproduce our results, and report our random search with weight-sharing results for each benchmark on two independent experimental runs.

Photo of Ameet Talwalkar

Ameet Talwalkar

Carnegie Mellon University and Determined AI

Ameet Talwalkar is co-founder and chief scientist at Determined AI and an assistant professor in the Machine Learning Department at Carnegie Mellon University. His primary interests are in the field of statistical machine learning, including problems at the intersection of systems and learning. He helped to create the SysML conference, led the initial development of the MLlib project in Apache Spark, is the coauthor of the graduate-level textbook Foundations of Machine Learning (MIT Press), and teaches an award-winning MOOC called Distributed Machine Learning with Apache Spark (edX).

Leave a Comment or Question

Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?

Join the conversation here (requires login)