Eloquence team submits multilingual ASR research for MLC-SLM challenge

Researchers from the Eloquence team have submitted their work for the MLC-SLM challenge, focusing on multilingual conversational speech recognition. Their submission explores three distinct methods: evaluating the baseline model with different projectors, utilizing the SLAM-ASR framework for a custom multilingual projector, and investigating the impact of contrastive learning and extended conversational context on recognition robustness. The goal is to advance speech language model architectures for real-world conversational data. AI

IMPACT Advances research in multilingual conversational speech recognition, potentially improving dialogue systems.

RANK_REASON Submission of a research paper to a challenge/workshop. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• alphaXiv
• arXiv
• CatalyzeX
• DagsHub
• Eloquence
• Gotit.pub
• Hugging Face
• Lorenzo Concina
• MLC-SLM Challenge
• ScienceCast
• SLAM-ASR

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →