UCT CS Research Document Archive

Improving the Transcription of Academic Lectures for Information Retrieval

Mbogho, Audrey and Stephen Marquard (2013) Improving the Transcription of Academic Lectures for Information Retrieval. The 12th International Conference on Machine Learning and Applications (ICMLA'13), Miami, FL, USA.

Full text available as:
PDF - Requires Adobe Acrobat Reader or other PDF viewer.

Abstract

Recording university lectures through lecture capture systems is increasingly common, generating large amounts of audio and video data. Transcribing recordings greatly enhances their usefulness by making them easy to search. However, the number of recordings accumulates rapidly, rendering manual transcription impractical. Automatic transcription, on the other hand, suffers from low levels of accuracy, partly due to the special language of academic disciplines, which standard language models do not cover. This paper looks into the use of Wikipedia to dynamically adapt language models for scholarly speech. We propose Ranked Word Correct Rate as a new metric better aligned with the goals of improving transcript searchability and specialist word recognition. The study shows that, while overall transcription accuracy may remain low, targeted language modelling can substantially improve searchability, an important goal in its own right.

EPrint Type:Conference Proceedings
Keywords:language model, speech recognition, lecture recording
Subjects:K Computing Milieux: K.3 COMPUTERS AND EDUCATION
I Computing Methodologies: I.7 DOCUMENT AND TEXT PROCESSING
ID Code:905
Deposited By:Mbogho, Audrey J. W.
Deposited On:28 October 2013