OpenAI trains neural network to play Minecraft and craft diamond tools with Video PreTraining
Original title:Learning to play Minecraft with Video PreTraining
AI summary
OpenAI trained a neural network on large amounts of unlabeled human Minecraft gameplay videos through Video PreTraining (VPT), using only a small amount of contractor-labeled data. After fine-tuning, the model can craft diamond tools, a task that typically takes a skilled human over 20 minutes and about 24,000 actions. The model uses the native human interface of keyboard and mouse, a step toward general-purpose computer-operating agents.
Key Takeaways
- OpenAI trained a neural network on large amounts of unlabeled human Minecraft gameplay videos…
- After fine-tuning,the model can craft diamond tools,a task that typically takes a skilled human…
- The model uses the native human interface of keyboard and mouse,a step toward general-purpose…
Quick Facts
- Published
- Updated
- Source
- OpenAI:官网动态(RSS · 排除企业/客户案例)
- Category
- Papers & Research
- AI score
- 62
原文说明 VPT 用少量标注数据配合海量无标注视频训练模型,读者可了解这种半监督路线对训练通用智能体的参考价值。
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.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: OpenAI trains neural network to play Minecraft and craft diamond tools with Video PreTraining. June 23, 2022. https://new.nat6.net/en/news/hhr7tugef77fyrld5og9tmfhu/