Low Resource Language
Low-resource language (LRL) research focuses on developing natural language processing (NLP) techniques for languages lacking substantial digital resources, aiming to bridge the technological gap between high- and low-resource languages. Current research emphasizes leveraging multilingual pre-trained models like Whisper and adapting them to LRLs through techniques such as weighted cross-entropy, data augmentation (including synthetic data generation), and model optimization methods like pruning and knowledge distillation. This work is crucial for promoting linguistic diversity, enabling access to technology for under-resourced communities, and advancing the broader field of NLP by addressing the challenges posed by data scarcity and linguistic variation.
Papers
Multi-Task Text Classification using Graph Convolutional Networks for Large-Scale Low Resource Language
Mounika Marreddy, Subba Reddy Oota, Lakshmi Sireesha Vakada, Venkata Charan Chinni, Radhika Mamidi
State-of-the-art in Open-domain Conversational AI: A Survey
Tosin Adewumi, Foteini Liwicki, Marcus Liwicki
Continuous Metric Learning For Transferable Speech Emotion Recognition and Embedding Across Low-resource Languages
Sneha Das, Nicklas Leander Lund, Nicole Nadine Lønfeldt, Anne Katrine Pagsberg, Line H. Clemmensen
Isomorphic Cross-lingual Embeddings for Low-Resource Languages
Sonal Sannigrahi, Jesse Read