OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments.
Tianbao XieDanyang ZhangJixuan ChenXiaochuan LiSiheng ZhaoRuisheng CaoToh Jing HuaZhoujun ChengDongchan ShinFangyu LeiYitao LiuYiheng XuShuyan ZhouSilvio SavareseCaiming XiongVictor ZhongTao YuPublished in: CoRR (2024)
Keyphrases
- open ended
- affect detection
- multi agent
- dynamic environments
- multi agent environments
- learning outcomes
- working environment
- multi agent systems
- highly dynamic
- autonomous agents
- intelligent agents
- robotic systems
- multiagent systems
- complex environments
- software agents
- cooperative
- multiple agents
- multi party
- multi modal
- learning experience
- mobile agents
- distributed intelligent
- autonomous entities
- multimedia
- inquiry learning
- exploratory search
- open systems
- computer assisted language learning
- agent model