Is It Novel and Why? Fine-Grained Patent Novelty Prediction Based on Passage Retrieval

2026-05-04

AI Summary

Current approaches to patent novelty assessment often oversimplify the task into claim-level binary classification, which is susceptible to spurious correlations and lacks fine-grained analysis. This work reframes novelty evaluation as a feature-level retrieval and reasoning task by decomposing patent claims, aligning them with specific passages in prior art documents, and identifying the distinctive features that confer novelty. To support this paradigm, we introduce FiNE-Patents, the first dataset comprising 3,658 annotated claims with feature-level prior art references, and propose an interpretable evaluation framework based on large language models (LLMs). Experimental results demonstrate that our approach significantly outperforms embedding-based baselines in both passage retrieval and identification of novel features, while exhibiting greater robustness against spurious correlations in claim-level classification.

Abstract

Novelty assessment is a critical yet complex task in the examination process for patent acceptance, requiring examiners to determine whether an invention is disclosed in a prior art document. The process involves intricate matching between specific features of a patent claim and passages in the prior art. While prior work has approached novelty prediction primarily as a binary classification task at the claim level, we argue that this formulation is susceptible to spurious correlations and lacks the granularity required for practical application. In this work, we introduce FiNE-Patents (Fine-grained Novelty Examination of Patents), a novel dataset comprising 3,658 first patent claims annotated with fine-grained, feature-level prior art references extracted from European Search Opinion (ESOP) documents. We propose shifting the evaluation paradigm from simple binary classification to a joint retrieval and abstract reasoning task at the feature level, requiring models to identify specific passages from a prior art document that disclose individual claim features, and to identify which features of a claim make it novel. We implement and evaluate LLM-based workflows that decompose claims into features, analyze each feature against prior art, and finally derive a claim-level novelty prediction. Our experiments demonstrate that these workflows outperform embedding-based baselines on passage retrieval and novel feature identification. Furthermore, we show that unlike trained classifiers, LLMs are robust against spurious correlations present in the claim-level novelty classification task. We release the dataset and code to foster further research into transparent and granular patent analysis.

Problem

• patent novelty
• fine-grained analysis
• prior art
• claim features
• spurious correlations

Innovation

• fine-grained novelty
• passage retrieval
• large language models
• patent analysis
• feature-level reasoning