A three-stage framework made up of neighborhood-aware retrieval, dual reranking, and score fusion that achieves the state of the art on average across five widely-used benchmarks and remains efficient at inference.
Abstract
Biomedical entity linking grounds mentions in clinical and scientific text to entities in a curated knowledge base (KB) with ontological structure, which supports downstream applications such as literature-scale information extraction and patient-record normalization. The task has several challenges at once: the KB contains large numbers of entities, mentions are often ambiguous, and gold labels follow annotation conventions specific to each corpus. To address these challenges, we propose PILOT, a three-stage framework made up of neighborhood-aware retrieval, dual reranking, and score fusion. The retriever injects ontological structure from both the query and KB side, by reformulating mentions and pooling entity embeddings. The retrieved pool is then scored from two complementary views, one over surface forms and one over context, and fused together. PILOT achieves the state of the art on average across five widely-used benchmarks and remains efficient at inference.
Biomedical named entity recognition (NER) and relation extraction (RE) remain challenging because biomedical texts contain ambiguous abbreviations, complex entity boundaries, domain-specific terminology, and implicit relations. This study proposes a prompt-enhanced and QLoRA-adapted large language model framework for biomedical information extraction. For NER, abbreviation-aware prompting supports candidate detection, contextual interpretation, boundary-aware generation, and schema-constrained outputs. For RE, entity markers identify a predefined target pair, while filtered UMLS and MeSH concepts provide concise evidence. DeepSeek-R1-Distill-Qwen-7B is adapted using LoRA over a 4-bit quantized frozen backbone. Experiments cover three NER and three RE datasets. Across three training seeds, the dataset-level macro-average F1 values are 0.909 ± 0.001 for NER and 0.787 ± 0.001 for RE. Seed-balanced paired bootstrap resampling with 10,000 aligned instance-level resamples confirms significant improvements over a matched deterministic simple-prompt baseline on all six datasets after Holm–Bonferroni correction (adjusted p < 0.001), with absolute F1 gains from +0.091 to +0.131. Repeated-run ablations show low variability and complementary contributions from task-structured prompting, knowledge filtering, deterministic validation, and parameter-efficient adaptation.
A dynamic parsing and relationship inference model for entity features that incorporates gating units and pooling strategies to dynamically update hidden states is proposed that outperforms the baseline model Adaptive Thresholding and Localized Context Pooling, particularly in extracting potential relationships across sentence–boundary entity pairs.
Haoxiang Yang, Guang-Hui Chang, Kuo Chen et al.· 0 citations
Scientific domain entity linking (EL) differs from general domain EL because mentions and entity names often lack lexical overlap. Another challenge is that specialized terminology is used in the scientific domain, which is rarely encountered in models pretrained on general domains. Therefore, models trained on general domains transfer poorly to scientific domains. To address this, in-domain fine-tuning is the natural remedy. However, many scientific domains lack expert-annotated data, motivating the need for a zero-human-annotation approach. Existing zero-shot methods heavily rely on LLMs to generate aliases across entire mention corpora, which incurs substantial computational cost, and those methods provide no mechanism to filter out noise from LLMs. To address these challenges, we propose Sci-ZSEL, a framework that selectively generates entity aliases with an LLM to control computational cost, and applies an ontology-aware filter to remove aliases that semantically drift toward ontology neighbors. Then, filtered aliases are used to construct pseudo-labeled mention-entity pairs for fine-tuning. To enable evaluation of EL under low lexical overlap, we also release a new animal science EL benchmark linked to three livestock trait ontologies, where mentions and entities exhibit substantially lower lexical overlap than in existing benchmarks. Across five benchmarks, Sci-ZSEL outperforms the non-fine-tuned baseline, is most useful on nonoverlapping mentions, and combining it with curated synonyms gives the best performance in most settings.
It is found that the triple composition and the training direction have the greatest impact, whereas the text encoder and hard-negatives sampling matter little and simple choices win: projecting text into the KG space with a linear head over concatenated subject, predicate, and object embeddings performs best.
Artem Bisliouk, E. Nosova, Heiko Paulheim et al.· 0 citations
Results demonstrate that InfoFlowEX equips LLMs with robust adaptability, achieving consistent gains over baselines with minimal task-specific customization, highlighting InfoFlowEX for real-world biomedical applications.
Wuyang Lan, Siqi Zhang, Wenzheng Wang et al.· Cell Reports Medicine· 0 citations
This paper addresses the placement of concepts that are absent from the target ontology—the out-of-knowledge-base setting—in which a textual mention must be assigned one or more insertion positions in the subsumption hierarchy rather than linked to an existing node.
Zhanna B. Sadirmekova, M. Sambetbayeva, B. Abdygalym et al.· Information· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
