Understanding Medical Conversations with Scattered Keyword Attention and Weak Supervision from Responses

Abstract

In this work, we consider the medical slot filling problem, i.e., the problem of converting medical queries into structured representations which is a challenging task. We analyze the effectiveness of two points: scattered keywords in user utterances and weak supervision with responses. We approach the medical slot filling as a multi-label classification problem with label-embedding attentive model to pay more attention to scattered medical keywords and learn the classification models by weak-supervision from responses. To evaluate the approaches, we annotate a medical slot filling data and collect a large scale unlabeled data. The experiments demonstrate that these two points are promising to improve the task.

MKA: A Scalable Medical Knowledge-Assisted Mechanism for Generative Models on Medical Conversation Tasks

“…In general, the medical dialogue methods can be divided into information retrieval-based methods and neural generative methods according to the types of the applied NLP techniques. The retrieval-based methods can be further classified into different subtypes, such as the entity inference [ 12 , 13 ], relation prediction [ 14 , 15 ], symptom matching and extraction [ 16 , 17 ], and slot filling [ 18 – 20 ]. However, the retrieval-based methods are not so intelligent and flexible that they required a well-defined user-built question and answer (Q&A) pool, which can offer different potential response to different kinds of answer.…”

ReMeDi: Resources for Multi-domain, Multi-service, Medical Dialogues

“…Most medical dialogue datasets contain only one domain (Wei et al 2018;Xu et al 2019;Shi et al 2020;Zhang et al 2020;Liu et al 2020;Wang, Song, and Xia 2018;Lin et al 2019) and/or one medical service (Wei et al 2018;Xu et al 2019;Liao et al 2020;Shi et al 2020;Lin et al 2021Lin et al , 2019. However, context information from other services and/or domains is often overlooked in a complete medical aid procedure.…”

“…Unlike common task-oriented dialogue systems (TDSs) for ticket or restaurant booking (Li et al 2017;Peng et al 2018;Wen et al 2017), MDSs are more challenging in that they require a great deal of expertise. For example, there are much more professional terms which are often expressed in colloquial language (Shi et al 2020).…”

“…Recently, an extensive effort has been made towards building data for MDS research (Liao et al 2020;Yang et al 2020;Shi et al 2020). However, they all have some limitations:…”

“…However, they only focus on one of the sub-tasks. For example, Shi et al (2020) study NLU, i.e., slot filling in medical consultation. Wei et al (2018) investigate DPL in medical diagnosis.…”