Knowledge-driven Subword Grammar Modeling for Automatic Speech Recognition in Tamil and Kannada

TL;DR

This paper introduces a knowledge-driven subword grammar modeling approach for ASR in Tamil and Kannada, improving recognition accuracy by capturing complex word formations and using linguistic knowledge.

Contribution

It presents a novel subword grammar weighted finite state transducer framework combined with heuristic segmentation and linguistic knowledge for agglutinative languages.

Findings

Achieved 12.39% and 13.56% word error rate reduction in Tamil and Kannada ASR systems.

Developed subword dictionaries and SG-WFST graphs based on language-specific morphological rules.

Enhanced recognition of complex words through heuristic segmentation and linguistic knowledge integration.

Abstract

In this paper, we present specially designed automatic speech recognition (ASR) systems for the highly agglutinative and inflective languages of Tamil and Kannada that can recognize unlimited vocabulary of words. We use subwords as the basic lexical units for recognition and construct subword grammar weighted finite state transducer (SG-WFST) graphs for word segmentation that captures most of the complex word formation rules of the languages. We have identified the following category of words (i) verbs, (ii) nouns, (ii) pronouns, and (iv) numbers. The prefix, infix and suffix lists of subwords are created for each of these categories and are used to design the SG-WFST graphs. We also present a heuristic segmentation algorithm that can even segment exceptional words that do not follow the rules encapsulated in the SG-WFST graph. Most of the data-driven subword dictionary creation algorithms are computation driven, and hence do not guarantee morpheme-like units and so we have used the linguistic knowledge of the languages and manually created the subword dictionaries and the graphs. Finally, we train a deep neural network acoustic model and combine it with the pronunciation lexicon of the subword dictionary and the SG-WFST graph to build the subword-ASR systems. Since the subword-ASR produces subword sequences as output for a given test speech, we post-process its output to get the final word sequence, so that the actual number of words that can be recognized is much higher. Upon experimenting the subword-ASR system with the IISc-MILE Tamil and Kannada ASR corpora, we observe an absolute word error rate reduction of 12.39% and 13.56% over the baseline word-based ASR systems for Tamil and Kannada, respectively.