Indian Language Resources–Evaluation Subcommittee Report

This report summarizes the findings and recommendations of the evaluation subcommittee of the IEEE prestandardization effort on standards for Indian language resources and evaluation for speech and language technology. Evaluation is a crucial aspect of any system. There are several generic challenges for the evaluation of speech and language technology, especially due to the ambiguous nature of human languages, as well as the interactive nature of many of these systems. Furthermore, there are additional challenges specific to Indian languages and the India context. This report summarizes the current metrics and datasets used for the evaluation of various technologies and identifies gaps that need to be addressed. Since there were dedicated subcommittees on speech, text, script, and accessibility, therefore, this report does not go into the details of various evaluation issues specific to these subareas. Instead, here the more generic issues that cut across these subdomains are discussed. A case study is also presented on machine translation (MT).


Sign up for our monthly newsletter to learn about new developments, including resources, insights and more.