Recurrent neural network models for disease name recognition using domain invariant features

Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, pages 2216–2225, Berlin, Germany, August 7-12, 2016. c⃝2016 Association for Computational Linguistics Recurrent neural network models for disease name recognition using domain invariant features Sunil Kumar Sahu and Ashish Anand Department of Computer Science and Engineering Indian Institute of Technology Guwahati Assam, India - 781039 {sunil.sahu, anand.ashish}@iitg.ernet.in Abstract Hand-crafted features based on linguistic and domain-knowledge play crucial role in determining the performance of disease name recognition systems. Such methods are further limited by the scope of these features or in other words, their ability to cover the contexts or word dependencies within a sentence. In this work, we focus on reducing such dependencies and pro- pose a domain-invariant framework for the disease name recognition task. In particu- lar, we propose various end-to-end recur- rent neural network (RNN) models for the tasks of disease name recognition and their classification into four pre-defined cate- gories. We also utilize convolution neu- ral network (CNN) in cascade of RNN to get character-based embedded features and employ it with word-embedded fea- tures in our model. We compare our mod- els with the state-of-the-art results for the two tasks on NCBI disease dataset. Our results for the disease mention recogni- tion task indicate that state-of-the-art per- formance can be obtained without relying on feature engineering. Further the pro- posed models obtained improved perfor- mance on the classification task of disease names. 1 Introduction Automatic recognition of disease names biomedical and clinical texts is of utmost impor-in tance for development of more sophisticated NLP systems such as information extraction, question answering, text summarization and so on (Rosario and Hearst, 2004). Complicate and inconsistent terminologies, ambiguities caused by use of ab- breviations and acronyms, new disease names, multiple names (possibly of varying number of words) for the same disease, complicated syntac- tic structure referring to multiple related names or entities are some of the major reasons for mak- ing automatic identification of the task difficult and challenging (Leaman et al., 2009). State-of- the-art disease name recognition systems (Mah- bub Chowdhury and Lavelli, 2010; Do˘gan and Lu, 2012; Dogan et al., 2014) depends on user defined features which in turn try to capture context keep- ing in mind above mentioned challenges. Feature engineering not only requires linguistic as well as domain insight but also is time consuming and is corpus dependent. Recently window based neural network ap- proach of (Collobert and Weston, 2008; Collobert et al., 2011) got lot of attention in different se- quence tagging tasks in NLP. It gave state-of-art results in many sequence labeling problems with- out using many hand designed or manually engi- neered features. One major drawback of this ap- proach is its inability to capture features from out- side window. Consider a sentence “Given that the skin of these adult mice also exhibits signs of de novo hair-follicle morphogenesis, we won- dered whether human pilomatricomas might orig- inate from hair matrix cells and whether they might possess beta-catenin-stabilizing mutations” (taken verbatim …

Content truncated to 500 words