WordAlchemy: A transformer-based Reverse Dictionary

TL;DR

This paper introduces WordAlchemy, a transformer-based reverse dictionary system supporting Indian languages, utilizing a novel TLM approach with mT5 to improve upon existing models.

Contribution

It presents a new cross-lingual reverse dictionary system for Indian languages using a transformer model with TLM, addressing limitations of previous methods.

Findings

Supports multiple Indian languages

Uses TLM with mT5 for better performance

Open-source and accessible system

Abstract

A reverse dictionary takes a target word's description as input and returns the words that fit the description. Reverse Dictionaries are useful for new language learners, anomia patients, and for solving common tip-of-the-tongue problems (lethologica). Currently, there does not exist any Reverse Dictionary provider with support for any Indian Language. We present a novel open-source cross-lingual reverse dictionary system with support for Indian languages. In this paper, we propose a transformer-based deep learning approach to tackle the limitations faced by the existing systems using the mT5 model. This architecture uses the Translation Language Modeling (TLM) technique, rather than the conventional BERT's Masked Language Modeling (MLM) technique.