The Compute Divide in Machine Learning: A Threat to Academic Contribution and Scrutiny?

TL;DR

This paper investigates how disparities in computing resources between industry and academia influence AI research, potentially limiting academic contributions, scrutiny, and the diffusion of advanced models, and proposes solutions to mitigate these effects.

Contribution

It provides a data-driven analysis of the compute divide's impact on academic research focus, scrutiny, and the shift towards industry-developed models, and suggests policy recommendations.

Findings

Academic research is less involved in compute-intensive topics like foundation models.

There is a shift towards open source, industry-developed pre-trained models in academia.

Reduced academic scrutiny may impact the evaluation and safety of influential models.

Abstract

There are pronounced differences in the extent to which industrial and academic AI labs use computing resources. We provide a data-driven survey of the role of the compute divide in shaping machine learning research. We show that a compute divide has coincided with a reduced representation of academic-only research teams in compute intensive research topics, especially foundation models. We argue that, academia will likely play a smaller role in advancing the associated techniques, providing critical evaluation and scrutiny, and in the diffusion of such models. Concurrent with this change in research focus, there is a noticeable shift in academic research towards embracing open source, pre-trained models developed within the industry. To address the challenges arising from this trend, especially reduced scrutiny of influential models, we recommend approaches aimed at thoughtfully expanding academic insights. Nationally-sponsored computing infrastructure coupled with open science initiatives could judiciously boost academic compute access, prioritizing research on interpretability, safety and security. Structured access programs and third-party auditing may also allow measured external evaluation of industry systems.