cs.CLOct 2, 2026

Enhancing Biomedical Named Entity Recognition via Multiple Programming Languages Instruction Tuning and Ensemble Method

Authors: Songtao Li, Yijia Zhang, Jianyuan Yuan, Shidi Zhang, Fengyu Zhang, Hongfei Lin

Organizations: School of Information Science and Technology, Dalian Maritime University, Dalian, China · School of Automation, Beijing Institute of Technology, Beijing, China · School of Computer Science and Engineering, Northeastern University, Shenyang, China · School of Computer Science and Technology, Dalian University of Technology, Dalian, China

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Instruction tuning has become a common paradigm for applying large language models (LLMs) to biomedical named entity recognition (BioNER). However, existing instruction-tuning approaches still face two key challenges. First, conventional natural-language instructions typically serialize BioNER annotations as flat textual outputs, providing limited structural constraints for typed entity extraction. Second, high-quality biomedical annotations are limited, and learning from a single serialized output form may restrict structural diversity and reduce model robustness. Although external biomedical knowledge can be introduced to alleviate data scarcity, it often requires costly resource construction. To address these challenges, we propose MITE, a Multiple Programming Languages Instruction Tuning and Ensemble method for BioNER. MITE reformulates BioNER as a structure-to-structure generation task by representing both instructions and entity outputs in code-formatted representations. Specifically, each training instance is transformed into multiple programming-language formats, including Python, C++, and Java, while preserving the same underlying entity semantics. These language-specific representations provide structurally diverse supervision without requiring external biomedical knowledge or additional annotations. During inference, MITE aggregates predictions from different code formats through an entity-level voting strategy, reducing language-specific prediction variance and improving robustness. Experiments on three widely used BioNER datasets demonstrate that MITE achieves competitive performance against representative BERT-based and LLM-based baselines, attaining the best results on two benchmarks. Ablation studies further validate the effectiveness of multi-language representations.

Figures & tables

Fig. 1: Examples of performing structured BioNER task with NL and Code instructions. In contrast to prompting with NL instructions, we utilize structured Code instructions to mitigate the output discrepancy between the pretraining and inference stages.
Fig. 2: Overview of the MITE framework. MITE converts each BioNER instance into code-formatted instructions in multiple programming languages, fine-tunes an LLM on the resulting multi-language training corpus, and aggregates language-specific predictions through entity-level voting to obtain the final BioNER prediction.
Dataset NameTrain / Dev / TestTypeEntities
BC5CDR-Chemical4559 / 4580 / 4796Chemical15828
BC2GM12632 / 2531 / 5065Gene24453
NCBI5432 / 923 / 942Disease6861
TABLE I: Statistics of the BioNER datasets used in our experiments.
MethodBackboneNCBI (Disease)BC2GM (Gene)BC5CDR (Chemical)
BERT-based Methods
BioBERTBioBERT88.1281.9483.83
PubMedBERTPubMedBERT86.9283.3693.20
AIONERBioBERT89.59–92.84
LLM-based Methods
UniversalNERLLaMA286.9682.42–
TABLE II: Main experimental results on the NCBI-Disease, BC2GM-Gene, and BC5CDR-Chemical datasets, evaluated using Micro-F1. The best and second-best results on each dataset are highlighted in bold and underlined , respectively.
Training SettingAvg. Score
MITE w/o Python{}_{\text{w/o Python}}88.42 (-0.30)
MITE w/o C++{}_{\text{w/o C++}}88.31 (-0.41)
MITE w/o Java{}_{\text{w/o Java}}88.56 (-0.16)
MITE88.72
TABLE III: Ablation of individual programming-language representations. Values in parentheses denote differences from the full MITE model.
Training SettingAvg. Score
Python 1×88.11 (-0.61)
C++ 1×88.03 (-0.69)
Java 1×88.21 (-0.51)
MITE sampled-train{}_{\text{sampled-train}}88.32 (-0.40)
Python 3×88.47 (-0.25)
C++ 3×88.38 (-0.34)
TABLE IV: Effect of data scale and programming-language diversity on average Micro-F1. Values in parentheses denote differences from the full MITE model.
More in this topic

Named Entity Recognition

Named Entity Recognition identifies and classifies people, places, organizations, and other entities in text, speech, and multimodal data.

See all 110 papers

Explore similar work

CardsList
Oct 2, 2026

A Guideline-Augmented Multi-Agent Framework for Schema-as-Code Biomedical Named Entity Recognition

Large language models (LLMs) have shown promising potential for biomedical named entity recognition (BioNER) through instruction following and in-context learning. However, existing LLM-based BioNER methods still face two key limitations. First, retrieved demonstrations and external biomedical knowledge provide limited support for dataset-specific annotation semantics, leaving entity boundaries, type scopes, and annotation conventions ambiguous. Second, free-form generation lacks sufficient structural control, often leading to invalid formats, hallucinated mentions, duplicated entities, and boundary errors. To address these limitations, we propose GAMA, a guideline-augmented multi-agent framework for schema-as-code BioNER. GAMA first induces candidate annotation rules from labeled training instances and verifies them against annotated data to construct reliable dataset-specific guideline memory. Guided by these verified rules, a planning component generates ranked span-type hypotheses with rationales, and a coding component converts them into schema-constrained entity objects. A verification module then checks span grounding, type validity, and structural compliance, and performs dual-loop refinement to correct invalid or low-confidence predictions. Experiments on three widely used BioNER datasets using two different LLMs demonstrate the effectiveness of GAMA, achieving the best performance in four out of six LLM-dataset combinations. Ablation studies further validate the contributions of its key components.

Songtao Li, Shidi Zhang, Jianyuan Yuan +1

School of Information Science and Technology, Dalian Maritime University, Dalian, China · School of Automation, Beijing Institute of Technology, Beijing, China · School of Computer Science and Engineering, Northeastern University, Shenyang, China

Named Entity RecognitionMulti-Agent LLM SystemsSchema-Constrained Generation
AskPDF
Mar 27, 2026

Analysing Lightweight Large Language Models for Biomedical Named Entity Recognition on Diverse Ouput Formats

Despite their strong linguistic capabilities, Large Language Models (LLMs) are computationally demanding and require substantial resources for fine-tuning, which is unadapted to privacy and budget constraints of many healthcare settings. To address this, we present an experimental analysis focused on Biomedical Named Entity Recognition using lightweight LLMs, we evaluate the impact of different output formats on model performance. The results reveal that lightweight LLMs can achieve competitive performance compared to the larger models, highlighting their potential as lightweight yet effective alternatives for biomedical information extraction. Our analysis shows that instruction tuning over many distinct formats does not improve performance, but identifies several format consistently associated with better performance.

Pierre Epron, Adrien Coulet, Mehwish Alam

HeKA | U1346, DIG · Inria, Inserm, Université Paris Cité, HeKA, UMR 1346, Paris, France · Télécom Paris, Institut Polytechnique de Paris, France +2

Named Entity RecognitionLLM Information ExtractionLLM EvaluationSmall Language ModelsClinical NLP
AskPDF
Oct 10, 2026

Training-Free BioNEL-RAG: Advancing Biomedical Entity Linking through Retrieval-Augmented Generation and Large Language Model

Biomedical Named Entity Linking (BioNEL) is the task of mapping mentions extracted from biomedical texts to canonical entities or concepts in the standardized knowledge base. Although existing methods have achieved remarkable performance in mapping accuracy, most rely on training with domain-specific datasets, which fundamentally limits their cross-domain adaptability. To address this limitation, we propose BioNEL-RAG, a novel training-free framework that combines retrieval-augmented generation (RAG) with large language models (LLMs). Within this framework, the RAG module focuses on retrieving candidate entities, while the large language model module plays the role of selecting the most appropriate candidate for entity normalization, which is based on the analysis of the mention's context, with the aid of its remarkable text comprehension capabilities. Experimental results demonstrate that BioNEL-RAG achieves performance comparable to existing methods without requiring training on subdomain datasets, whereas these methods necessitate fine-tuning on specific subdomains such as genes, diseases, or chemical compounds.

Ziyi Chen, Hui Liu, Ziyang Ling +2

School of Biomedical Engineering, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei, China · Department of Precision Machinery and Precision Instruments, University of Science and Technology of China, Hefei, China

Biomedical Entity LinkingEntity Linking
AskPDF