Shangeth Rajaa

Latest

• DualTurn: Learning Turn-Taking from Dual-Channel Generative Speech Pretraining
• SpeechLLM: Multi-Modal LLM for Speech Understanding
• Improving spoken language identification with MAP-Mix
• Improving end-to-end spoken language understanding with prosodic attention and distillation
• Skit-S2I: An Indian Accented Speech to Intent Dataset
• Learning speaker representation with semi-supervised learning approach for speaker profiling
• RL based framework to generate optimal data quality remediation sequence for machine learning
• Towards Automated Deep Learning: Analysis of the AutoDL challenge series 2019