Back to feed
arXiv cs.CL
arXiv cs.CL
7/8/2026
NAVER LABS System Re-implementation for the IWSLT 2026 Instruction-Following Task

NAVER LABS System Re-implementation for the IWSLT 2026 Instruction-Following Task

Short summary

NAVER LABS re-implements their IWSLT 2025 instruction-following pipeline for the IWSLT 2026 Shared Task, using SeamlessM4T-v2-large as speech encoder and Qwen3-4B-Instruct as the LLM backbone. The system retains a three-stage approach (projector alignment, text-only LoRA pre-training, multimodal merging) and adds 100k synthetic instruction-following examples across ten speech-centric task types. The primary model achieves COMET 0.781 on EN-ZH speech translation and BERTScore-F1 0.346 on English SQA on the MCIF benchmark.

  • •Re-implements NAVER LABS IWSLT 2025 pipeline for IWSLT 2026 using SeamlessM4T-v2-large and Qwen3-4B-Instruct
  • •Adds 100k synthetic instruction-following examples across 10 speech-centric task types
  • •Achieves COMET 0.781 on EN-ZH speech translation and BERTScore-F1 0.346 on English SQA

Generated with AI, which can make mistakes.

#research-breakthrough#ai-tools
Read full article at arXiv cs.CL

Is this a good recommendation for you?

Comments

Failed to load comments. Please try again.

Explore more

MLLP-VRAIN UPV system for the IWSLT 2026 Simultaneous Speech Translation task
arXiv cs.CL
Do LLMs Need Architectural Changes for Simultaneous Speech Translation? A Prefix-to-Prefix Data Driven Approach
arXiv cs.CL
LLM-INSTRUCT at UZH Shared Task 2026: Constraint-Aware Retrieval and Selective Debate for Paragraph-Level Argument Mining
arXiv cs.CL
Efficiently Adapting Spoken Language Models for the Singaporean Context
arXiv cs.CL