Word similarity datasets created for six major Indian languages
Word Similarity Datasets for Indian Languages: Annotation and Baseline Systems
Computation and LanguageMachine Learning
Summary
Word similarity tasks help computers understand how related different words are, which improves language technologies. The authors built new word similarity datasets for six widely spoken Indian languages by translating and adjusting English datasets. They also tested some modern computer language models using these datasets and shared baseline results. This work provides new tools for better language understanding in these Indian languages.
What this means in practice
- •For language technology developers: Evaluate and improve word representation models for Indian languages using newly available word similarity datasets.
- •For machine translation teams: Use the datasets to better assess semantic similarity in Indian languages and improve translation quality.
Authors
Syed S. Akhtar, Arihant Gupta, Avijit Vajpayee, Arjit Srivastava, M. Shrivastava
Abstract
With the advent of word representations, word similarity tasks are becoming increasing popular as an evaluation metric for the quality of the representations. In this paper, we present manually annotated monolingual word similarity datasets of six Indian languages - Urdu, Telugu, Marathi, Punjabi, Tamil and Gujarati. These languages are most spoken Indian languages worldwide after Hindi and Bengali. For the construction of these datasets, our approach relies on translation and re-annotation of word similarity datasets of English. We also present baseline scores for word representation models using state-of-the-art techniques for Urdu, Telugu and Marathi by evaluating them on newly created word similarity datasets.