Search papers, labs, and topics across Lattice.
NAVER LABS Europe submitted an advanced system for instruction-following speech processing at IWSLT 2026, achieving notable improvements in ASR, ST, and SQA tasks from English to Chinese, Italian, and German. By replacing the traditional speech projector with SpeechMapper, which learns a speech-to-LLM embedding using only ASR data, and introducing a synthetic dataset called fakACL for scientific presentations, the team enhanced their multi-stage training pipeline. This approach enabled them to outperform last year's top system while utilizing a more compact model with a weaker LLM backbone, ultimately tying for first place in the overall short track ranking.
A novel speech-to-LLM embedding projector and synthetic dataset allowed NAVER LABS to outperform last year's champion system with a more compact architecture.
In this paper, we describe NAVER LABS Europe's submission to the instruction-following speech processing short track at IWSLT 2026. We participate again in the constrained setting, developing systems capable of jointly performing ASR, ST, and SQA from English speech into Chinese, Italian, and German. Building on our previous submission, ranked first in last year's short track, we update our multi-stage training pipeline by replacing the speech projector with SpeechMapper, a method for learning a speech-to-LLM embedding projector using only ASR data. In addition, we introduce a synthetic SQA dataset, fakACL, composed of artificially generated scientific presentations. This dataset is built by prompting the LLM backbone, segmenting the generated talks, and synthesizing speech with SeamlessM4T-large-v2. The combination of an improved speech projection mechanism and domain-specific synthetic data allows our model to outperform last year's best short-track system, while being considerably more compact and relying on a weaker LLM backbone. This year's results place our system tied for first place in the overall short track ranking.
