scienceworld-target-identifier

To locate objects in simulated environments, this Skill processes observation data and filters items based on predefined taxonomy and contextual relevance to match your target description. It ranks candidates by accessibility and proximity for efficient retrieval.

Object detection in environment exploration works by processing room observations to filter and rank observable items. It matches target descriptions like "living thing" or "container" against taxonomy references to identify relevant objects.

Yes, you need classification scripts and taxonomy references to accurately identify items in room observations. These dependencies allow the Skill to parse observation data and match target descriptions within the simulated environment.

The best way to find and prioritize a target object is to run this Skill with your observation and target description. It filters matching items and ranks them based on accessibility, proximity, and task relevance for retrieval.

Yes, you can filter observable items in ScienceWorld using taxonomy filtering. The Skill processes observation data to match descriptive criteria against taxonomy references, enabling targeted environmental analysis and object retrieval.