Computer Science

Power-law scaling to assist with key challenges in artificial intelligence

Y. Meir, S. Sardi, et al.

This study reveals how optimized test errors in deep learning diminish as database sizes increase, crucial for swift decision-making. Conducted by a team of researchers at Bar-Ilan University, this research sets a benchmark for assessing training complexity across various machine learning tasks and algorithms.

00:00

Playback language: English

Index

Abstract

This paper investigates power-law scaling in deep learning, specifically focusing on how optimized test errors converge to zero with increasing database size. The study explores this relationship for both single-epoch and multi-epoch training, examining the impact on test error and the implications for rapid decision-making. The research establishes a benchmark for measuring training complexity and provides a quantitative hierarchy of machine learning tasks and algorithms.

Publisher

Scientific Reports

Published On

Nov 12, 2020

Authors

Yuval Meir, Shira Sardi, Shiri Hodassman, Karin Kisos, Itamar Ben-Noam, Amir Goldental, Ido Kanter

Related Publications

Explore these studies to deepen your understanding of the subject.

Medicine and Health

What Is in There for Artificial Intelligence to Support Mental Health Care for Persons with Serious Mental Illness? Opportunities and Challenges

B. Wang, C. K. Grønvik, et al.

Psychology

What Is in There for Artificial Intelligence to Support Mental Health Care for Persons with Serious Mental Illness? Opportunities and Challenges

B. Wang, C. K. Grønvik, et al.

Business

A humanistic model of corporate social responsibility in e-commerce with high-tech support in the artificial intelligence economy

E. B. Zavyalova, V. A. Volokhina, et al.

Medicine and Health

Prospective multicenter study using artificial intelligence to improve dermoscopic melanoma diagnosis in patient care

Not Available

Listen, Learn & Level Up

Over 10,000 hours of research content in 25+ fields, available in 12+ languages.

No more digging through PDFs, just hit play and absorb the world's latest research in your language, on your time.

listen to research audio papers with researchbunny