Opinion

OpenAI on Dota 2 results: self-play took the system from below human to superhuman

Original title:More on Dota 2

AI summary

OpenAI said its Dota 2 results show that, with enough compute, self-play can take machine learning systems from far below human level to superhuman level. The system progressed within a month from barely matching high-ranked players to beating top professionals, and has continued to improve since; data in self-play systems improves automatically as agents get stronger, free from the quality ceiling of supervised deep learning training datasets.

Key Takeaways

  • OpenAI said its Dota 2 results show that,with enough compute,self-play can take machine learning…
  • The system progressed within a month from barely matching high-ranked players to beating top…
  • data in self-play systems improves automatically as agents get stronger,free from the quality…

Quick Facts

Published
Updated
Source
OpenAI:官网动态(RSS · 排除企业/客户案例)
Category
Opinion
AI score
71
Read the original Read the full article at the original source
Why it matters

OpenAI 借 Dota 2 结果说明自我博弈为何能突破监督学习的数据上限,对理解强化学习路线有参考价值。

Related entities

Source & copyright

This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.

Original article · AIHOT entry

How to Cite

When citing this article, keep the title, date, and permalink, and credit Orange News as the source.

Orange News: OpenAI on Dota 2 results: self-play took the system from below human to superhuman. August 16, 2017. https://new.nat6.net/en/news/jj17vhggylfu9uqetb3oni4ob/

View all →