Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Jul 12, 2026

Decoding Natural Behavior from Neuroethological Embedding

Decoding Natural Behavior from Neuroethological Embedding

Published on: October 3, 2025

BIOSMILE: a semantic role labeling system for biomedical verbs using a maximum-entropy model with automatically

Richard Tzong-Han Tsai¹, Wen-Chi Chou, Ying-Shan Su

¹Institute of Information Science, Academia Sinica, Nankang, Taipei 115, Taiwan, PRoC. thtsai@iis.sinica.edu.tw

BMC Bioinformatics

|September 4, 2007

Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

PSIP1 Orchestrates Immune Escape in Osteosarcoma: Insights From Single-Cell Analysis and Implications for Immunotherapy Resistance.

International journal of genomics·2026

Same author

Evaluating real-world deployment of an HL7-CDA-aligned LLM for ICD-10-CM coding.

NPJ digital medicine·2026

Same author

'Interaction annealing' to determine effective quantized valence and orbital structure: an illustration with ferro-orbital order in WTe<sub>2</sub>.

npj computational materials·2026

Same author

Attention-PestNet: hierarchical scaled dot-product attention for insect pest detection.

Plant methods·2025

Same author

Leveraging large language models for the deidentification and temporal normalization of sensitive health information in electronic health records.

NPJ digital medicine·2025

Same author

Blind Testing Method to Validate the Efficiency of ICD-10-CM Artificial Intelligence Coding Modules.

Studies in health technology and informatics·2025

Developing a biomedical semantic role labeling (SRL) system using a specialized proposition bank significantly improves the extraction of crucial biomedical relations, outperforming general-purpose systems.

Area of Science:

Biomedical Informatics
Computational Linguistics
Natural Language Processing

Background:

Bioinformatics tools aid biomedical research by processing literature for large-scale experiments.
Information extraction (IE) systems using NLP are vital in biomedicine, particularly for relation extraction (e.g., protein-protein, gene-disease).
Existing systems often overlook essential phrases (adverbial, prepositional) that describe relation details like location, manner, and timing.

Purpose of the Study:

To develop BIOSMILE, a biomedical semantic role labeling (SRL) system for enhanced biomedical relation extraction.
To leverage a specialized biomedical proposition bank (BioProp) for training the SRL system.
To improve the capture of detailed relational information often missed by standard IE systems.

Main Methods:

Related Experiment Videos

Last Updated: Jul 12, 2026

Decoding Natural Behavior from Neuroethological Embedding

Decoding Natural Behavior from Neuroethological Embedding

Published on: October 3, 2025

Constructed BIOSMILE, a biomedical SRL system employing a maximum entropy (ME) machine-learning model.
Trained BIOSMILE on BioProp, a semi-automatically annotated biomedical proposition bank.
Focused on 30 key biomedical verbs central to describing molecular events.

Main Results:

Training on BioProp improved the SRL system's F-score by 21.45% compared to a system trained on newswire corpora.
Incorporating automatically generated template features further boosted the overall F-score by 0.52%.
Specific argument classifications (ArgM-LOC, ArgM-MNR, Arg2) showed statistically significant performance gains.

Conclusions:

A biomedical-specific proposition bank is essential for effective SRL in the biomedical domain.
Biomedical and newswire sentence characteristics, cross-domain framesets, and verb usage variations impact SRL performance.
Template features incorporating words, NE types, and POS tags significantly improved classification accuracy for adjunct arguments, crucial for biomedical SRL.