TARSphere benchmarks long-horizon robots in changing physical and social scenes
Original title: TARSphere: A Robot-Oriented Full-Stack Framework for Long-Horizon Embodied Agents in Evolving Physical and Social Environments
The benchmark has 360 episodes across 30 indoor scenes, and tests of eight MLLMs found updates and human interaction still hard.
TARSphere is a full-stack framework for robots that must follow physical and social constraints as environments and human needs change. Its simulator runs policies with embodied NPCs and constrained dialogue, and an event-driven state machine updates task state from completed actions. A frozen multimodal model plans through executable skills, hierarchical memory, and execution feedback. TARSphere-Bench has 360 episodes in 30 indoor scenes; an evaluation of eight models found general-purpose models could beat embodied ones, while task updates and human interaction remained difficult.
- Five conditions cover perception, disruption, task updates, coordination, and intent.
- Tasks evaluated are navigate-and-report and fetch-and-deliver.
- Collisions and falls are cited as reasons to score execution safety with task success.
From the original article
1 min read at huggingface.co
Robots must act under physical and social constraints while adapting to changes in their environment and human needs. Existing embodied-agent systems often focus on specified goals or abstract execution into symbolic actions, limiting evaluation in an evolving world. We introduce TARSphere, a robot-oriented full-stack framework for physically and socially grounded long-horizon embodied agents.
By Yicheng Xu, Zhikai Xu, Ziyuan Chen · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.