PatEx-Gym benchmarks agents on patent prior-art search
Original title: Novel or Not? PatEx-Gym: Agentic Prior-Art Search for Novelty Assessment in Patent Examination
A 4B PatEx-Agent trained on synthetic trajectories is presented as competitive with much larger general-purpose agents.
PatEx-Gym is a framework for training and evaluating agents that search prior art and draft novelty assessments for patent examination. PatEx-Bench supplies 1K examination tasks, retrieval corpora of more than 1M patents and 52K scientific papers, and a runtime for iterative search, document visits, reasoning, and Office Action generation. PASS synthesizes annotation-free search trajectories, PatEx-SynData, using document visits as relevance signals. A 4B PatEx-Agent fine-tuned on those trajectories is reported as competitive with substantially larger general-purpose agents, while prior-art discovery remains hard even for strong frontier models.
Unconfirmed: The abstract does not name the frontier models or report their exact scores.
- PatEx-Bench contains 1K examination tasks.
- Corpora exceed 1M patents and include 52K scientific papers.
- PASS builds trajectories without manual annotation.
- PatEx-Agent is trained at 4B scale by supervised fine-tuning.
- Decision accuracy alone did not fully reflect examination quality.
From the original article
1 min read at huggingface.co
Patent examination is a challenging real-world task that requires agents to discover sparse prior art, iteratively refine queries, inspect technical evidence, and produce novelty assessments documented in Office Actions (OAs) across heterogeneous scientific and technical domains.
By Qiyao Wang, Xinyi Chen, Hongbo Wang · 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.