Work

The Eloquence submission for Task 2 of the Interspeech 2026 MLC-SLM challenge

Year: 2026

Type: conference-paper

Abstract: This paper details the Eloquence team's approach to Task 2 of the 2nd MLC-SLM challenge at Interspeech 2026, which involves multilingual Multiple-Choice Question Answering (MCQA) across 21 languages.Three approaches are explored.First, we fine-tune Voxtral-Mini-3B via LoRA with cross-lingual data augmentation, ASR transcript augmentation and timestamp-aware audio cropping, achieving 0.72 macroaccuracy on evaluation Phase 2. Second, we apply multimodal in-context learning (ICL) to the frozen Voxtral-24B model to correct a strong label bias, reaching 0.81, our best result.Third, a training-free retrieval system based on a three-layer voiceanchored memory combining acoustic identity, semantic content, and a knowledge graph achieves 0.68.All three systems substantially outperform the official baseline.

Source: Workshop on Multilingual Conversational Speech Language Model (MLC-SLM)

DOI: 10.21437/mlcslm.2026-11

Authors: Jordi Luque, Lorenzo Concina, Marco Matassoni, Alessio Brutti, Filippo Vella

Institutions: Telefónica Innovación Digital (Spain), Fondazione Bruno Kessler, National Research Council

Language: English

Topic: Speech Recognition and Synthesis

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: bronze

Funders: European High Performance Computing Joint Undertaking, European Commission, HORIZON EUROPE Framework Programme

Award: Multilingual and Cross-cultural interactions for context-aware, and bias-controlled dialogue systems for safety-critical applications