Home » world » Achieving Human-like Intelligence in AI: Exploring Computational Rules, Brain Activity, and Number Sense

Achieving Human-like Intelligence in AI: Exploring Computational Rules, Brain Activity, and Number Sense

2023-06-23 09:59:40

Nikkei Science August 2023 issue

Editorial department


While ChatGPT can solve high school-level math problems, it also makes rudimentary arithmetic mistakes that make it hard to imagine someone with the same intelligence. Why? It is also a question of how the human brain solves mathematics. The latest research has clarified what computational rules AI learns from text alone, and what brain activity is like when humans solve computational problems. It is also becoming clear that many animals, including humans, have a “number sense.” Even in the same mathematics, ChatGPT is completely wrong when it comes to puzzles that require flexible thinking. New research has also begun with the aim of realizing AI that can inspire, discover, and invent.

Human intelligence reflected by ChatGPT
Masaaki Demura Cooperation: Ryota Kanai/Ippei Fujisawa/Tomoya Nakai
Number Sense: Born or Learned?J. Beck / S. Clark
You can beat ChatGPT with this math puzzle
M. Bartels

1687521194
#Special #Feature #Mathematical #Brain #Nikkei #Science

You may also like

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Adblock Detected

Please support us by disabling your AdBlocker extension from your browsers for our website.