Research agent
Ask about “Potential-Based Reward Shaping”
The research agent plans a search, reads the papers it finds, and writes an answer that cites every paper-specific claim, so you can open the source and check it. Threads are saved to your account, so you can come back, follow up, and export them as Markdown.
Scoped to what you ask
Start from a paper, a topic, or one of your Library collections, or search across everything in the index.
Grounded in the papers
It works from indexed metadata and abstracts, and reads arXiv full-text passages and extracted figures when a question needs more evidence.
Every claim links back
Answers cite the records they rely on, and each run shows its sources and the steps it took.
Questions you could ask
- What are the most important recent directions in Potential-Based Reward Shaping?
- Which approaches currently compete within Potential-Based Reward Shaping?
- Where is the evidence uncertain or contested in Potential-Based Reward Shaping?