Transfer language space with similar domain adaptation: a case study with hepatocellular carcinoma

Journal of Biomedical Semantics

Omar Kallas¹  Scott Jeffery Lee¹  Patricia Balthazar¹  Terry Desser²  Daniel Rubin³  Amara Tariq⁴  Judy Wawira Gichoya⁵  Imon Banerjee⁵ 

[1] Department of Radiology, Emory University, Atlanta, GA, USA;Department of Radiology, Stanford University, Palo Alto, CA, USA;Department of Radiology, Stanford University, Palo Alto, CA, USA;Department of Biomedical Data Science, Stanford University, Palo Alto, CA, USA;Machine Intelligence in Medicine and Imaging (MI ∙2) Lab, Mayo Clinic, Phoenix, AZ, USA;Machine Intelligence in Medicine and Imaging (MI ∙2) Lab, Mayo Clinic, Phoenix, AZ, USA;Department of Radiology, Emory University, Atlanta, GA, USA;

关键词: Transfer learning;    Language model;    Radiology report;    BERT;    Word2vec;   

DOI  :  10.1186/s13326-022-00262-8

来源: Springer

【 摘 要 】

BackgroundTransfer learning is a common practice in image classification with deep learning where the available data is often limited for training a complex model with millions of parameters. However, transferring language models requires special attention since cross-domain vocabularies (e.g. between two different modalities MR and US) do not always overlap as the pixel intensity range overlaps mostly for images.MethodWe present a concept of similar domain adaptation where we transfer inter-institutional language models (context-dependent and context-independent) between two different modalities (ultrasound and MRI) to capture liver abnormalities.ResultsWe use MR and US screening exam reports for hepatocellular carcinoma as the use-case and apply the transfer language space strategy to automatically label imaging exams with and without structured template with > 0.9 average f1-score.ConclusionWe conclude that transfer learning along with fine-tuning the discriminative model is often more effective for performing shared targeted tasks than the training for a language space from scratch.

【 授权许可】

CC BY