Fine-tuning large neural language models for biomedical natural language processing
Open accessThe DOI matches its Crossref record; 16 of the 16 references it cites were found in a registry or an index.
Abstract
Large neural language models have transformed modern natural language processing (NLP) applications. However, fine-tuning such models for specific tasks remains challenging as model size increases, especially with small labeled datasets, which are common in biomedical NLP. We conduct a systematic study on fine-tuning stability in biomedical NLP. We show that fine-tuning performance may be sensitive to pretraining settings and conduct an exploration of techniques for addressing fine-tuning instability. We show that these techniques can substantially improve fine-tuning performance for low-resource biomedical NLP applications. Specifically, freezing lower layers is helpful for standard BERT- B A S E models, while layerwise decay is more effective for BERT- L A R G E and ELECTRA models. For low-resource text similarity tasks, such as BIOSSES, reinitializing the top layers is the optimal strategy. Overall, domain-specific vocabulary and pretraining facilitate robust models for fine-tuning. Based on these findings, we establish a new state of the art on a wide range of biomedical NLP applications.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
RoBERTa: A Robustly Optimized BERT Pretraining Approach
Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., Levy, O., Lewis, M., Zettlemoyer, L., & Stoyanov, V. (2019) · arXiv · doi:10.48550/arxiv.1907.11692 · Cited by 16,515
BioBERT: a pre-trained biomedical language representation model for biomedical text mining
Lee, J., Yoon, W., Kim, S., Kim, D., Kim, S., So, C. H., & Kang, J. (2020) · Bioinformatics · doi:10.1093/bioinformatics/btz682 · Cited by 7,218
GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding
Wang, A., Singh, A., Michael, J., Hill, F., Levy, O., & Bowman, S. (2018) · Proceedings of the 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP · doi:10.18653/v1/w18-5446 · Cited by 4,222
Know What You Don't Know: Unanswerable Questions for SQuAD
Rajpurkar, P., Jia, R., & Liang, P. (2018) · arXiv · doi:10.48550/arxiv.1806.03822 · Cited by 2,242
Know What You Don’t Know: Unanswerable Questions for SQuAD
Rajpurkar, P., Jia, R., & Liang, P. (2018) · Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers) · doi:10.18653/v1/p18-2124 · Cited by 2,240
Domain-specific language model pretraining for biomedical natural language processing
Gu, Y., Tinn, R., Cheng, H., Lucas, M., Usuyama, N., Liu, X., Naumann, T., Gao, J., & Poon, H. (2022) · ACM Transactions on Computing for Healthcare · doi:10.1145/3458754 · Cited by 2,170 · cited 5×
“We ground our study in BLURB, a recently-proposed comprehensive benchmark for biomedical NLP comprising 6 tasks and 13 datasets (Gu et al., 2021).”
BackgroundMethodInfluentialSelf-citationAlsentzer, E., Murphy, J., Boag, W., Weng, W.-H., Jindi, D., Naumann, T., & McDermott, M. (2019) · Proceedings of the 2nd Clinical Natural Language Processing Workshop · doi:10.18653/v1/w19-1909 · Cited by 1,736 · cited 1×
“Recently, there has been increasing interest in biomedical pretraining (Lee et al. , 2020b; Peng et al. , 2019; Alsentzer et al. , 2019; Gu et al. , 2021) and applications (Bressem et al. , 2020; Trieu et al. , 2020).”
BackgroundSelf-citationMaking Pre-trained Language Models Better Few-shot Learners
Gao, T., Fisch, A., & Chen, D. (2021) · Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) · doi:10.18653/v1/2021.acl-long.295 · Cited by 1,380 · cited 1×
“Adversarial training can also help instability issues (Jiang et al. , 2020; Zhu et al. , 2020; Cheng et al. , 2021) and prompt-based learning has been shown to work well in low-resource settings (Schick and Schütze, 2021; Gao et al. , 2021).”
BackgroundSuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems
Wang, A., Pruksachatkun, Y., Nangia, N., Singh, A., Michael, J., Hill, F., Levy, O., & Bowman, S. R. (2019) · arXiv · doi:10.48550/arxiv.1905.00537 · Cited by 1,051
Neural Machine Translation of Rare Words with Subword Units
Sennrich, R., Haddow, B., & Birch, A. (2016) · Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/p16-1162 · Cited by 974 · cited 1×
“To address the problem of out-of-vocabulary words, neural language models generate a vocabulary from subword units, using Byte-Pair Encoding (BPE; Sennrich et al., 2016) or variants such as WordPiece (Kudo and Richardson, 2018).”
Method
Cited by
Chen, Q., Hu, Y., Peng, X., Xie, Q., Jin, Q., Gilson, A., Singer, M. B., Ai, X., Lai, P.-T., Wang, Z., et al. (2025) · Nature Communications · doi:10.1038/s41467-025-56989-2 · Cited by 219
“In the biomedical domain, a primary challenge is the limited availability of labeled datasets, which have a signi fi cantly lower scale than those in the general domain (e.g., a biomedical sentence similarity dataset only has 100 labeled instances in total 31 ) 32,33 .”
InfluentialLarge Language Models are Few-shot Testers: Exploring LLM-based General Bug Reproduction
Kang, S., Yoon, J., & Yoo, S. (2023) · 2023 IEEE/ACM 45th International Conference on Software Engineering (ICSE) · doi:10.1109/icse48619.2023.00194 · Cited by 204
Data extraction methods for systematic review (semi)automation: a living systematic review
Schmidt, L., Olorisade, B. K., McGuinness, L. A., Thomas, J., & Higgins, J. P. T. (2021) · F1000Research · doi:10.12688/f1000research.51117.1 · Cited by 106
Kim, J., Wang, K., Weng, C., & Liu, C. (2024) · The American Journal of Human Genetics · doi:10.1016/j.ajhg.2024.08.010 · Cited by 53
Clay, B., Bergman, H. I., Salim, S., Pergola, G., Shalhoub, J., & Davies, A. H. (2025) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2025.109808 · Cited by 41
Data extraction methods for systematic review (semi)automation: Update of a living systematic review
Schmidt, L., Finnerty Mutlu, A. N., Elmore, R., Olorisade, B. K., Thomas, J., & Higgins, J. P. T. (2025) · F1000Research · doi:10.12688/f1000research.51117.3 · Cited by 27
BigBIO: A Framework for Data-Centric Biomedical Natural Language Processing
Fries, J. A., Weber, L., Seelam, N., Altay, G., Datta, D., Garda, S., Kang, M., Su, R., Kusa, W., Cahyawijaya, S., et al. (2022) · arXiv · doi:10.48550/arxiv.2206.15076 · Cited by 17
ProteinGLUE multi-task benchmark suite for self-supervised protein modeling
Capel, H., Weiler, R., Dijkstra, M., Vleugels, R., Bloem, P., & Feenstra, K. A. (2022) · Scientific Reports · doi:10.1038/s41598-022-19608-4 · Cited by 11
Xian, R. P., Lee, A. J., Lolla, S., Wang, V., Cui, Q., Ro, R., & Abbasi-Asl, R. (2024) · arXiv · doi:10.48550/arxiv.2402.10527
Metadata
- Authors
- Tinn, R., Cheng, H., Gu, Y., Usuyama, N., Liu, X., Naumann, T., Gao, J., & Poon, H.
- Venue
- Patterns
- Publisher
- Elsevier BV
- Published
- 2023-04
- Volume
- 4
- Issue
- 4
- Pages
- 100729
- ISSN
- 2666-3899 (ISSN-L)
- OpenAlex
- W4365511667
- Cited by
- 169
- References
- 16
- FWCI
- 13.44
- Type
- journal article
- Language
- en
- Open access
- gold
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.