Sign In

Communications of the ACM

ACM TechNews

Opening the Black Box

View as: Print Mobile App Share:
Opening a black box.

Black box explorations are designed to process information without revealing any details about their reasoning.


Researchers at Arizona State University (ASU) and the University of California, Los Angeles hope to enable scientists and processor designers to understand the underlying reasoning of deep learning accelerator designs through explainable-design space exploration (DSE).

ASU's Shail Dave said hardware and software designs are typically optimized via black box mechanisms that "require excessive amounts of trial runs because of their lack of explainability and reasoning involved in how selecting a design configuration affects the design's overall quality."

Explainable-DSE simplifies the accelerator's decision-making process so choices of design methods can be made in minutes rather than days or weeks, supporting smaller, more systematic, and more energy-efficient models.

Dave's algorithm can investigate design solutions relating to multiple applications, including those differing in functionality or processing traits, while resolving their product execution inefficiencies.

From ASU News
View Full Article


Abstracts Copyright © 2023 SmithBucklin, Washington, D.C., USA


No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account