Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a configurable semantic chunking framework that enhances biomedical information extraction in retrieval-augmented generation (RAG) systems by addressing the limitations of fixed-size chunking, which often fragments semantic evidence. The new approach integrates various techniques such as entity-preserving windows and tiered trigger prioritization, allowing for more effective extraction of relevant information. Evaluation on benchmarks like GM-CIHT and DDI shows a significant F1 score improvement from 74.2% to 82.6%, demonstrating the framework's efficacy in optimizing biomedical data extraction processes.
Configurable semantic chunking boosts F1 scores by over 8 points in biomedical extraction tasks, revealing the limitations of fixed-size approaches.
BioMedRAG introduced retrieval-augmented generation with a learned chunk scorer for biomedical information extraction. However, it relies on fixed-size chunking which can fragment semantic evidence. We propose a configurable semantic chunking framework that addresses this limitation by combining entity-preserving windows, trigger-centered chunking, proposition-first extraction, tiered trigger prioritization, and hierarchical relation resolution. The framework integrates with BioMedRAG by replacing only the chunk construction stage while preserving the embedding model, learned chunk scorer, generator, and evaluation protocol. We evaluate the framework on biomedical relation extraction benchmarks (GM-CIHT, DDI, ChemProt) and adverse event classification (ADE). On GM-CIHT, the full hybrid configuration achieves 82.6% F1, improving over the fixed-size baseline (74.2% F1) by 8.4 points under our experimental setup. Cross-dataset analysis shows that semantic chunking improves extraction datasets with explicit relation cues, such as GM-CIHT and DDI, while fixed chunking remains competitive or stronger for dense biochemical extraction and binary classification settings such as ChemProt and ADE. By externalizing chunking logic into configuration files, the framework provides an interpretable and adaptable alternative to rigid fixed-size chunking for biomedical RAG pipelines.