We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Domain-Relevant Embeddings for Medical Question Similarity

Abstract: The rate at which medical questions are asked online far exceeds the capacity of qualified people to answer them, and many of these questions are not unique. Identifying same-question pairs could enable questions to be answered more effectively. While many research efforts have focused on the problem of general question similarity for non-medical applications, these approaches do not generalize well to the medical domain, where medical expertise is often required to determine semantic similarity. In this paper, we show how a semi-supervised approach of pre-training a neural network on medical question-answer pairs is a particularly useful intermediate task for the ultimate goal of determining medical question similarity. While other pre-training tasks yield an accuracy below 78.7% on this task, our model achieves an accuracy of 82.6% with the same number of training examples, and an accuracy of 80.0% with a much smaller training set.
Comments: Machine Learning for Health (ML4H) at NeurIPS 2019 - Extended Abstract
Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL); Machine Learning (stat.ML)
Cite as: arXiv:1910.04192 [cs.LG]
  (or arXiv:1910.04192v2 [cs.LG] for this version)

Submission history

From: Clara McCreery [view email]
[v1] Wed, 9 Oct 2019 18:19:48 GMT (856kb,D)
[v2] Fri, 15 Nov 2019 01:06:47 GMT (622kb,D)

Link back to: arXiv, form interface, contact.