Intent Identification and Entity Extraction for Healthcare Queries in Indic Languages

Findings of the Association for Computational Linguistics: EACL 2023, pages 1870–1881 May 2-6, 2023 ©2023 Association for Computational Linguistics Intent Identification and Entity Extraction for Healthcare Queries in Indic Languages 1Ankan Mullick 2Ishani Mondal∗ 1Sourjyadip Ray∗ 3R Raghav 1G Sai Chaitanya and 1Pawan Goyal {1ankanm, 1sourjyadipray}@kgpian.iitkgp.ac.in, 2imondal@umd.edu 3rraghavr@cs.cmu.edu, {1gajulasai@, 1pawang@cse.}iitkgp.ac.in 1Computer Science and Engineering Department, IIT Kharagpur, India 2University of Maryland, College Park, USA 3School of Computer Science, Carnegie Mellon University, USA Abstract Scarcity of data and technological limitations for resource-poor languages in developing countries like India poses a threat to the de- velopment of sophisticated NLU systems for healthcare. To assess the current status of vari- ous state-of-the-art language models in health- care, this paper studies the problem by initially proposing two different Healthcare datasets, Indian Healthcare Query Intent-WebMD and 1mg (IHQID-WebMD and IHQID-1mg) and one real world Indian hospital query data in English and multiple Indic languages (Hindi, Bengali, Tamil, Telugu, Marathi and Gujarati) which are annotated with the query intents as well as entities. Our aim is to detect query in- tents and extract corresponding entities. We perform extensive experiments on a set of mod- els in various realistic settings and explore two scenarios based on the access to English data only (less costly) and access to target language data (more expensive). We analyze context spe- cific practical relevancy through empirical anal- ysis. The results, expressed in terms of overall F1 score show that our approach is practically useful to identify intents and entities. 1 Introduction Healthcare is a top priority for every country. Peo- ple across the world ask millions of health-related queries, hoping to get a response from a domain expert (Gebbia et al., 2020). These queries mostly deal with medical history of patients, possible drug interactions, disease related concerns, treatment protocols and so on. Conversational agents for healthcare play a pivotal role by facilitating useful information dissemination (Li et al., 2020; Maniou and Veglis, 2020). In order to understand these ∗Authors contributed equally queries better, practical conversational systems for healthcare need to be developed. However, the pri- mary obstacle in developing such technologies for low-resource languages is the lack of usable data (Mehta et al., 2020; Daniel et al., 2019; Liu, 2022). India is a country with a diverse language speak- ing population suffering from abject poverty and low-economic status (Mohanty, 2010; Pande and Yazbeck, 2003). This linguistic diversity and com- plex socio-economic situation in India certainly poses significant challenges in developing auto- matic healthcare systems; and there is a lack of linguistic resources specific to the medical domain. For example, situations such as the patient and the doctor speaking in different languages, is not an un- common situation in rural India. These individuals are unable to avail the existing systems and facil- ities which exist mainly in the English language. Recent efforts in developing automatic translation systems, even from extremely low resource lan- guages such as ‘Mundari’ and ‘Gondi’ (Joshi et al., 2019), should ideally improve this situation, but there is no extensive …

Content truncated to 500 words