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

Download:

Current browse context:

cs.LG

Change to browse by:

cs

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: Understanding Medical Conversations: Rich Transcription, Confidence Scores & Information Extraction

Abstract: In this paper, we describe novel components for extracting clinically relevant information from medical conversations which will be available as Google APIs. We describe a transformer-based Recurrent Neural Network Transducer (RNN-T) model tailored for long-form audio, which can produce rich transcriptions including speaker segmentation, speaker role labeling, punctuation and capitalization. On a representative test set, we compare performance of RNN-T models with different encoders, units and streaming constraints. Our transformer-based streaming model performs at about 20% WER on the ASR task, 6% WDER on the diarization task, 43% SER on periods, 52% SER on commas, 43% SER on question marks and 30% SER on capitalization. Our recognizer is paired with a confidence model that utilizes both acoustic and lexical features from the recognizer. The model performs at about 0.37 NCE. Finally, we describe a RNN-T based tagging model. The performance of the model depends on the ontologies, with F-scores of 0.90 for medications, 0.76 for symptoms, 0.75 for conditions, 0.76 for diagnosis, and 0.61 for treatments. While there is still room for improvement, our results suggest that these models are sufficiently accurate for practical applications.
Subjects: Machine Learning (cs.LG)
Cite as: arXiv:2104.02219 [cs.LG]
  (or arXiv:2104.02219v1 [cs.LG] for this version)

Submission history

From: Izhak Shafran [view email]
[v1] Tue, 6 Apr 2021 01:16:59 GMT (99kb,D)

Link back to: arXiv, form interface, contact.