about
The Next Decade in AI (2020) (arxiv.org)
3 points by layer8 on May 5, 2022 | hide | past | pdf | discuss on HN

In plain words: A proposed AI design pairs general-purpose learning with built-in knowledge and step-by-step reasoning, organized around models of how people think. Unlike today's push for ever-larger training sets and more computing power, this mix is argued to yield AI that is richer and more robust.

Abstract · The Next Decade in AI: Four Steps Towards Robust Artificial Intelligence

Recent research in artificial intelligence and machine learning has largely emphasized general-purpose learning and ever-larger training sets and more and more compute. In contrast, I propose a hybrid, knowledge-driven, reasoning-based approach, centered around cognitive models, that could provide the substrate for a richer, more robust AI than is currently possible.

Gary Marcus
arXiv:2002.06177 · cs.AI, cs.LG · submitted Feb 14, 2020 · updated Feb 19, 2020
abstract · pdf · 5 figures

add comment on HN
Also discussed: Nov 2024 (2 points, 1 comment) · Jul 2024 (1 point, 0 comments) · Aug 2020 (1 point, 0 comments) · Jun 2020 (3 points, 1 comment) · Feb 2020 (2 points, 0 comments) · Feb 2020 (1 point, 0 comments) · Feb 2020 (2 points, 0 comments) · Feb 2020 (6 points, 0 comments)