Shashi Kumar, Esaú Villatoro-Tello, Sergio Burdisso, Kadri Hacioglu, Thibault Bañeras-Roux, Hasindri Watawana, Dairazalia Sanchez-Cortes, Srikanth Madikeri, Petr Motlicek, Andreas Stolcke
Standard LLM-based speech recognition systems typically process utterances in isolation, limiting their ability to leverage conversational context. In this work, we study whether multimodal context from prior turns improves LLM-based ASR and how to represent that context efficiently. We find that, after supervised multi-turn training, conversational context mainly helps with the recognition of contextual entities. However, conditioning on raw context is expensive because the prior-turn audio token sequence grows rapidly with conversation length. To address this, we propose Abstract Compression, which replaces the audio portion of prior turns with a fixed number of learned latent tokens while retaining corresponding transcripts explicitly. On both in-domain and out-of-domain test sets, the compressed model recovers part of the gains of raw-context conditioning with a smaller prior-turn audio footprint. We also provide targeted analyses of the compression setup and its trade-offs.
Thibault Bañeras-Roux, Sergio Burdisso, Esaú Villatoro-Tello, Dairazalia Sánchez-Cortés, Shiran Liu, Severin Baroudi, Shashi Kumar, Hasindri Watawana, Manjunath K E, Kadri Hacioglu, Petr Motlicek, Andreas Stolcke
Conventional end-to-end automatic speech recognition (ASR) systems rely on paired speech-text data for domain adaptation. Recent LLM-based ASR architectures connect a speech encoder to a large language model via a projection module, enabling adaptation with text-only data. However, this introduces a modality gap, as the LLM is not exposed to the noisy representations produced by the speech projector. We investigate whether small amounts of speech can mitigate this mismatch. We compare three strategies: text-only adaptation, paired speech-text adaptation, and mixed batching (MB), which combines both. Experiments in in-domain and out-of-domain settings show that even limited speech consistently improves performance. Notably, MB using only 10% of the target-domain (less than 4 hours) speech achieves word error rates comparable to, or better than, conventional ASR fine-tuning with the full dataset, indicating that small amounts of speech provide a strong modality-alignment signal.
Yanis Labrak, David Grünert, Séverin Baroudi, Jiyun Chun, Pawel Cyrta, Sergio Burdisso, Ahmed Hassoon, David Liu, Adam Rothschild, Reed Van Deusen, Petr Motlicek, Andrew Perrault, Ricard Marxer, Thomas Schaaf
Long-context audio reasoning is underserved in both training data and evaluation. Existing benchmarks target short-context tasks, and the open-ended generation tasks most relevant to long-context reasoning pose well-known challenges for automatic evaluation. We propose a synthetic data generation pipeline designed to serve both as a training resource and as a controlled evaluation environment, and instantiate it for first-visit doctor-patient conversations with SOAP note generation as the task. The pipeline has three stages, persona-driven dialogue generation, multi-speaker audio synthesis with overlap/pause modeling, room acoustics, and sound events, and LLM-based reference SOAP note production, built entirely on open-weight models. We release 8,800 synthetic conversations with 1.3k hours of corresponding audio and reference notes. Evaluating current open-weight systems, we find that cascaded approaches still substantially outperform end-to-end models.
Shashi Kumar, Srikanth Madikeri, Juan Zuluaga-Gomez, Iuliia Thorbecke, Esaú Villatoro-Tello, Sergio Burdisso, Petr Motlicek, Karthik Pandia, Aravind Ganapathiraju
In traditional conversational intelligence from speech, a cascaded pipeline is used, involving tasks such as voice activity detection, diarization, transcription, and subsequent processing with different NLP models for tasks like semantic endpointing and named entity recognition (NER). Our paper introduces TokenVerse, a single Transducer-based model designed to handle multiple tasks. This is achieved by integrating task-specific tokens into the reference text during ASR model training, streamlining the inference and eliminating the need for separate NLP models. In addition to ASR, we conduct experiments on 3 different tasks: speaker change detection, endpointing, and NER. Our experiments on a public and a private dataset show that the proposed method improves ASR by up to 7.7% in relative WER while outperforming the cascaded pipeline approach in individual task performance. Our code is publicly available: https://github.com/idiap/tokenverse-unifying-speech-nlp
Iuliia Thorbecke, Juan Zuluaga-Gomez, Esaú Villatoro-Tello, Shashi Kumar, Pradeep Rangappa, Sergio Burdisso, Petr Motlicek, Karthik Pandia, Aravind Ganapathiraju
The training of automatic speech recognition (ASR) with little to no supervised data remains an open question. In this work, we demonstrate that streaming Transformer-Transducer (TT) models can be trained from scratch in consumer and accessible GPUs in their entirety with pseudo-labeled (PL) speech from foundational speech models (FSM). This allows training a robust ASR model just in one stage and does not require large data and computational budget compared to the two-step scenario with pre-training and fine-tuning. We perform a comprehensive ablation on different aspects of PL-based streaming TT models such as the impact of (1) shallow fusion of n-gram LMs, (2) contextual biasing with named entities, (3) chunk-wise decoding for low-latency streaming applications, and (4) TT overall performance as the function of the FSM size. Our results demonstrate that TT can be trained from scratch without supervised data, even with very noisy PLs. We validate the proposed framework on 6 languages from CommonVoice and propose multiple heuristics to filter out hallucinated PLs.
Séverin Baroudi, Yanis Labrak, Shashi Kumar, Joonas Kalda, Sergio Burdisso, Pawel Cyrta, Juan Ignacio Alvarez-Trejos, Petr Motlicek, Hervé Bredin, Ricard Marxer
Extracting patient medical conditions from code-switched clinical spoken dialogues is challenging due to rapid turn-taking and highly overlapped speech. We present a robust system evaluated on the DISPLACE-M dataset of real-world Hinglish medical conversations. We propose an End-to-End Neural Diarization with Vector Clustering approach (EEND-VC) to accurately resolve dense and speaker overlaps in Doctor-Patient Conversations (DoPaCo). For transcription, we adapt a Qwen3 ASR model via domain-specific fine-tuning, Devanagari script normalization, and dialogue-level LLM error correction, achieving an 18.59% tcpWER. We benchmark open and proprietary LLMs on medical condition extraction, comparing our text-based cascade system against a multimodal End-to-End (E2E) audio framework. While proprietary E2E models set the performance ceiling, our open cascaded architecture is highly competitive, as it achieved first place out of 25 participants in the DISPLACE-M challenge. All implementations are publicly released.
Thibault Bañeras-Roux, Shashi Kumar, Driss Khalil, Sergio Burdisso, Petr Motlicek, Shiran Liu, Mickael Rouvier, Jane Wottawa, Richard Dufour
Automatic Speech Recognition (ASR) is traditionally evaluated using Word Error Rate (WER), a metric that is insensitive to meaning. Embedding-based semantic metrics are better correlated with human perception, but decoder-based Large Language Models (LLMs) remain underexplored for this task. This paper evaluates their relevance through three approaches: (1) selecting the best hypothesis between two candidates, (2) computing semantic distance using generative embeddings, and (3) qualitative classification of errors. On the HATS dataset, the best LLMs achieve 92--94\% agreement with human annotators for hypothesis selection, compared to 63\% for WER, also outperforming semantic metrics. Embeddings from decoder-based LLMs show performance comparable to encoder models. Finally, LLMs offer a promising direction for interpretable and semantic ASR evaluation.