Speech
-
Decoding Naturalistic Speech from fMRI with Temporal Modeling and Multimodal Alignment
2026-10-15 15:17:15Corentin Caris, Louis Jalouzot, Nona Rajabi, G\'eraud Ilinca, Ahmed Hassayoune, Asilbek Muminov, Thadd\'ee Delebarre, Clara Fonteneau, Charles Maussion, Bertrand Thirion, Alexis Thual, Antoine CollasAbstract
Reconstructing what a person perceives from their brain activity is a stringent test of whether neural recordings carry meaningful information. Naturalistic speech offers a rich testbed for this challenge: it unfolds over multiple timescales, but fMRI's noisy, temporally smeared measurements make this structure hard to recover. We introduce NOEMA, a temporal convolutional decoder that aligns windows of consecutive fMRI measurements with multimodal stimulus embeddings. These embeddings combine complementary text, audio, and speech representations of the same stimulus segment, enabling its retrieval directly from brain activity. On two naturalistic speech datasets, Cephalonauts One and LeBel, NOEMA retrieves the correct segment among 2,000 candidates within the top 10 for more than 42% of held-out queries. This substantially outperforms baseline decoders, an advantage that persists when the candidate pool is extended with thousands of distractors absent from training. Controlled ablations establish the contributions of temporal modeling and complementary stimulus representations.
Abstract (translated)
URL
https://arxiv.org/abs/2610.22049
PDF
-
Frame-Wise Streaming for Efficient Learned Speech Compression
2026-10-15 14:42:23Yuhan Liu, Jun Xu, Fengxi Zhang, Zhengxue Cheng, Li SongAbstract
Speech compression is increasingly important in scenarios such as real-time and satellite communication. Neural codecs have shown superior Rate-Distortion (RD) performance over traditional signal-processing codecs. Models with long-range dependency encoders/decoders and entropy models outperform convolution-based structures, but they still lack streaming support since they rely on sequence-level correlations. In this paper, we propose StreamLSC, a frame-level streaming learned speech compression model built on three fundamental components of streaming networks, i.e. Boundary-Cached Causal Convolution, Intra-Frame Layer Normalization, and RNN-mode RWKV. Our model incorporates a streamable entropy quantizer module, ensuring efficient context model while preserving causal constraints. Experiments show that the streaming model achieves performance comparable to its offline counterpart, attaining state-of-the-art rate-distortion results, significant gains in subjective evaluations, -51.37% (ViSQOL) and -53.99% (PESQ) BD-RATE savings on average, and improvements of 0.2154 and 0.5156 in BD-ViSQOL and BD-PESQ.
Abstract (translated)
URL
https://arxiv.org/abs/2610.21998
PDF
-
Learning Speech Attribute Representations via Multi-Teacher Distillation
2026-10-15 14:15:32Maryem Bouziane, Salima Mdhaffar, Yannick Est\`eveAbstract
Recent post-training approaches have shown that self-supervised speech encoders can be adapted to learn a single utterance-level speech attribute through knowledge distillation. In this work, we investigate whether a single speech encoder can simultaneously preserve multiple complementary utterance-level speech representations. We propose a multi-attribute multi-teacher distillation framework that learns semantic, speaker, language, and emotion representations from four frozen teacher models using a shared SSL backbone and attribute-specific branches. Experiments on multilingual semantic retrieval, speaker verification, language identification, and speech emotion recognition show that the proposed model achieves performance close to dedicated single-task students while requiring only a single shared encoder. Furthermore, the learned layer aggregation strategies reveal that different speech attributes naturally emerge at different depths of the SSL encoder, providing new insights into the organization of speech representations.
Abstract (translated)
URL
https://arxiv.org/abs/2610.21941
PDF
-
Dynamic ordering in sequence generation in Transformers: experiments on low-resource Tagging scenarios
2026-10-15 13:36:46Giacomo Bais, Tejaswini DeoskarAbstract
This paper introduces TagInsert, a Transformer-based model that challenges the common left-to-right (L2R) generation paradigm in Transformers, for sequence labelling tasks where dependencies may not naturally align with a L2R order. TagInsert lifts the L2R constraint by removing the decoder-side attention mask and enabling the model to decide a preferred order of tag generation at inference time. We experiment on the common Part-of-Speech tagging and the more complex Combinatory Categorial Grammar (CCG) supertagging tasks on multiple languages and low-resource data settings. We show that, although encoder-only models like BERT remain strong performers for tagging tasks, TagInsert frequently outperforms a L2R encoder-decoder model, achieves results comparable to encoder-only models under data-scarce conditions, and shows better results on the long-tail of rare categories. Our analyses show that TagInsert alleviates error propagation typical of autoregressive models by learning to defer difficult predictions to later stages in the decoding process.
Abstract (translated)
URL
https://arxiv.org/abs/2610.21862
PDF
-
NICE Toolbox: Nonverbal Interpersonal Communication Exploration Toolbox
2026-10-15 12:41:38Aleksandr Evgrashin, Carolin Schmitt, Timo L\"ubbing, Ashutosh Jha, Buket Naz Zeren, Sophie K. Bauer, Johannes Kopf-Beck, Anton K. G. Marx, Anamaria Semm, G\"okce Erg\"un, Senya PolikovskyAbstract
Understanding human communication is a long-standing goal in psychological research. People express themselves through many verbal and nonverbal channels: posture, gesture, gaze, facial expression, and speech. Measurement and annotation of these signals has traditionally required time-consuming manual work by trained human labellers. The Nonverbal Interpersonal Communication Exploration (NICE) Toolbox is a free, open-source Python framework that automatically measures these signals by combining existing state-of-the-art computer vision and speech-processing methods into a single software package. It includes support for multi-camera systems, interchangeable detection algorithms, 2D and 3D visualization of detector outputs, evaluation on the user's custom data, and export to third-party software. With the release of this tool, we aim to make data exploration of interpersonal communication signals scalable, reproducible, and accessible for psychological research.
Abstract (translated)
URL
https://arxiv.org/abs/2610.21786
PDF
-
CF-AVSS: A Coarse-to-Fine Audio-Visual Speech Separation Framework with Target-Speaker Refinement and Post-Fusion for Real-World AVSE Challenge
2026-10-15 12:30:11Longjie Luo, Zhongshu Hou, Yuepeng Li, Gang Yang, Shidong Shang, Fan ZhouAbstract
Audio-visual speech enhancement (AVSE) leverages visual cues to recover target speech from mixtures corrupted by reverberation, noise, and overlapping speakers. We propose a coarse-to-fine two-stage AVSE framework. In the first stage, a progressive SPAV-TFGridNet is extended to two synchronized visual streams to produce visually ordered coarse estimates by progressively suppressing interfering speech while preserving target reverberation and shared background noise. In the second stage, a TFGridNet-based target speaker refinement network takes the original mixture and both coarse estimates as input to further recover target speech and improve signal quality. The coarse and refined estimates are finally combined using development-set-calibrated fusion. On the ISCSLP 2026 Real-World AVSE Challenge, our system tied for first place on Track 1 and ranked second on Track 2 in OVRL, while achieving the lowest CER on both tracks at 12.31% and 14.79%, respectively.
Abstract (translated)
URL
https://arxiv.org/abs/2610.21773
PDF
-
Full-Duplex Interaction with Speech, Expressions and Gestures
2026-10-15 07:54:33Siyin Wang, Wenyi Yu, Terumi Chiba, Xianzhao Chen, Xiaohai Tian, Jun Zhang, Lu Lu, Chao ZhangAbstract
Human face-to-face conversation requires coordinating speech, gestures and expressions across both speaking and listening. We present ELLSA 2, an end-to-end full-duplex dialogue model that jointly optimizes speech interaction and streaming gesture-expression generation. An autoregressive, chunk-wise synthesizer combines pretrained acoustic features with distinct listening and speaking hidden states from the dialogue backbone and speech synthesizer, enabling continuous non-verbal generation even when the model is verbally silent. To translate dialogue understanding into vivid semantic gestures, we introduce contextual semantic calls and template guidance, separating whether to invoke an available semantic gesture from how to realize it. We also establish a curated data and evaluation pipeline for dialog-aware co-speech gesture generation and investigate reference-based multimodal large language models (MLLMs) as automatic judges. Experiments show that ELLSA 2 remains highly competitive with open-source full-duplex speech interaction models and outperforms cascaded baselines across dialog-aware gesture and expression metrics. Semantic template guidance further significantly improves semantic gesture quality. Strong system-level correlations with human judgments support these reference-based scores as useful preliminary metrics for full-duplex and semantic gesture assessment. Together, these results advance dialogue modeling from spoken responses toward coordinated multimodal interaction.
Abstract (translated)
URL
https://arxiv.org/abs/2610.21452
PDF
-
SITAR: A Situational Distortion Metric for Automatic Speech Recognition Evaluation
2026-10-15 07:48:03Mingzhe Lu, Jinhao Liu, Yunpeng Li, Hang Su, Lichun Fan, Jian LuanAbstract
In automatic speech recognition (ASR), transcription errors with similar lexical or semantic discrepancies can have markedly different consequences for the situation conveyed by speech. Capturing these differences requires not merely contextual assessment, but an understanding of what the spoken discourse describes, establishes, or seeks to bring about. However, relying on LLM judges introduces computational cost and scoring variability. We introduce SITAR, a situational distortion metric for ASR, realized through a comprehensive end-to-end construction framework. Specifically, this framework aggregates situational severity and cumulative extent into a unified score via a bounded recursive formulation. Using targets derived from this formulation with limited offline LLM support, we train SITAR as a lightweight BERT-family regressor that scores reference-hypothesis pairs in a single forward pass, entirely free of any prohibitive runtime costs. Recognizing that introducing a new metric demands rigorous justification, we formalize a principled validation protocol spanning both extrinsic validity, assessed at macro and micro levels, and intrinsic scorer reliability. The former, evaluated across diverse datasets and models, demonstrates that SITAR reliably tracks broad capability scaling while providing orthogonal diagnostic utility, whereas the latter, examined on a purpose-built controlled benchmark, confirms the operational soundness of the scoring instrument itself. Through systematic validation, we demonstrate that SITAR holds strong potential as a practical, lightweight complement to existing ASR metrics. We will release the full recipes (code, model weights, dataset, etc.) after the internal review to support community research.
Abstract (translated)
URL
https://arxiv.org/abs/2610.21440
PDF
-
Automatic Sentence Completion Using Bengali Audio Speech with a Transformer-Based Early Speech Recognition System
2026-10-15 03:24:24Md Asif Akram, Arif Faisal, Sakir Hossain Faruque, Sharun Akter KhushbuAbstract
Dementia severely impairs communication abilities and turns everyday conversations into challenging experiences. Individuals affected by dementia frequently struggle to express their thoughts or complete sentences, making it difficult for family members and caregivers to understand their needs. This study proposes a novel approach to assist dementia patients by developing an integrated system that detects signs of dementia through Bangla speech and text analysis and provides sentence auto-completion support during conversations. To achieve this, recorded voices from individuals with and without dementia were analysed, and key acoustic features were extracted to capture vocal patterns associated with cognitive decline. A transformer-based model was employed to transcribe voice signals into text for linguistic analysis. In the classification task, several Machine Learning (ML) and Deep Learning (DL) models were evaluated using both acoustic and textual representations, and the LSTM model architectures demonstrated strong capabilities for modelling contextual patterns in speech and language data. For the sentence auto-completion task, a sequential learning model based on a Bidirectional LSTM (Bi-LSTM) network with an attention mechanism was developed to sequentially complete partially expressed Bangla sentences during conversation. This research bridges the gap in dementia analysis in an underrepresented language and implements a sentence auto-completion framework for cognitive disorders.
Abstract (translated)
URL
https://arxiv.org/abs/2610.21164
PDF
-
VoxTurn: Answering Several Turn-Taking Questions in One Pass with One Model
2026-10-15 03:06:58Yang XiaoAbstract
A cascaded voice agent must decide, throughout each silence of its user, whether the user has finished their turn: answering too early cuts the user off, answering too late makes the agent slow. It often needs related decisions too: will the user soon go on, is speech during its own turn an interruption or a backchannel, is a sound speech at all. With open detectors, each extra decision costs another model or another call, and on the 14-language eot-bench they cut users off in more than a quarter of mid-turn pauses. Because they are trained on segmented utterances, this mismatch is the largest single source of error. We present VoxTurn, a 0.8B speech LLM that casts these decisions as multiple-choice questions, scores the hidden state of each option, and answers any set of them in one forward pass with a block-isolating attention mask, exactly as if asked one at a time. Trained on real conversational pauses with labels derived from their timing, and re-queried while a silence lasts, VoxTurn reaches 11.5% false cut-offs in English at a 300 ms latency budget, comparable to two commercial detectors, and a 14-language mean of 19.9%, the lowest among open models, at 8.3 ms per decision. In Pipecat and LiveKit it cuts off 17% and 9% of turns, and it ranks fifth of 24 in TurnBench end-of-turn detection without any TurnBench training data. The model, code and framework plugins are released.
Abstract (translated)
URL
https://arxiv.org/abs/2610.21150
PDF
-
CoSTA: Compressed Speech Token Adaptation for Efficient Spoken Dialogue Models
2026-10-15 03:00:10Yingzhuo Deng, Chong Li, Bingxin Liu, Yi Shu, Tianyu Peng, Can Xie, Siyao Song, Shuo Ren, Jiajun ZhangAbstract
Spoken Dialogue Models (SDMs) typically share a speech-token space for generation and synthesis, leading to long autoregressive sequences and costly inference. Lower-rate speech tokenizers shorten sequences but may introduce compression-quality trade-offs and costly replacement without breaking this coupling. To address this limitation, we propose CoSTA, a compatibility-preserving post-hoc framework that decouples generation and synthesis token spaces while retaining the original speech tokenizer and decoder. CoSTA extends the generation token space with compressed representations via mean-initialized vocabulary extension, pronunciation adaptation, and dialogue adaptation, followed by deterministic expansion back to the original speech-token space before synthesis. On bilingual speech-dialogue benchmarks, CoSTA achieves approximately $1.7\times$ expanded speech-token generation speed, lower full-response latency, and higher throughput while preserving response quality. Analyses show more sustained interaction under limited budgets and a practical trade-off among compression, decoding efficiency, and quality. Further experiments demonstrate cross-architecture transfer and complementarity to low-rate tokenization.
Abstract (translated)
URL
https://arxiv.org/abs/2610.21143
PDF
-
Code-Switching Detection via Language-Selective Dimensions in Speech Foundation Models
2026-10-14 22:26:33Jie Chi, Zakaria Aldeneh, Masha Fedzechkina, Maureen de Seyssel, Natalie SchluterAbstract
Code-switching (CS), the alternation between languages within an utterance, is common in multilingual speech but underrepresented in pretraining corpora. Increasing CS coverage requires finding relevant utterances in large speech collections so that they can be included in training data. One way to find such utterances is to transcribe the speech and detect language alternation in the text, but this depends on obtaining accurate multilingual transcriptions. We propose a lightweight method that detects CS directly from speech. It uses frame-level language information in pretrained speech encoders to determine whether both languages are present within an utterance. Specifically, we identify language-selective hidden-state dimensions using monolingual recordings and aggregate their frame-level activations into a CS prediction. With a small CS set used to tune a decision rule, we show that the Whisper-small encoder achieves a mean F1 of $0.798$ across 14 language pairs in CS-FLEURS, compared with $0.776$ for a CNN and $0.751$ for a HuBERT linear baseline trained for monolingual language identification. Our findings suggest that pretrained speech encoders already encode the information needed to identify CS utterances, and they provide a lightweight tool for building more representative multilingual corpora.
Abstract (translated)
URL
https://arxiv.org/abs/2610.18726
PDF
-
Probing Neural Audio Codecs with Speaker Traits and Speech Characteristics
2026-10-14 21:27:37Jihwan Lee, Tiantian Feng, Minu Kim, Junhyeok Lee, Sean Foley, Thanathai Lertpetchpun, Najim Dehak, Dani Byrd, Shrikanth NarayananAbstract
Neural audio codecs are increasingly used for low-bitrate speech coding and serve as tokenizers for speech language models, yet it remains unclear how their codebook hierarchies encode speaker traits and speech characteristics. We systematically study 17 multi-codebook neural audio codecs across seven traits: sex, age, accent, voice quality, emotion, speech flow, and non-verbal sounds. For each codec, we reconstruct speech using partial codebooks and use machine listeners to compare trait predictions on each reconstruction. We measure prediction agreement with reweighted Cohen's \k{appa}, and define trait depth as a proxy for the codebook depth by which each trait is encoded and recoverable. Overall, sex has the shallowest trait depth, followed by speech flow, whereas voice quality and age group require deeper stacks; emotion, accent, and non-verbal sounds fall in between. Within each trait, speech-flow classes are at relatively similar trait depths, whereas emotion, accent, and non-verbal sounds differ more. Additionally, we explore how certain codec design choices relate to trait depths. These findings show that codebook hierarchies are trait-dependent and may not exhibit uniform coarse-to-fine structures, motivating trait-specific codec evaluation and design for future applications.
Abstract (translated)
URL
https://arxiv.org/abs/2610.20955
PDF
-
Task-Centric Multimodal Learning for Cognitive Impairment Detection
2026-10-14 20:34:38Sankalp Vaish, Rishank Singh, Sumukh Veeramalla, Caroline Summerour, John A. Batsis, Xiaohui LiangAbstract
Dementia affects millions of older adults worldwide, yet early detection remains difficult outside clinical settings. Traditional models primarily rely on text or voice-based responses, capturing only a limited view of human cognition despite the inherently multimodal nature of human expression. This work proposes a contactless, multimodal screening framework that leverages video, audio, and text signals collected during structured voice-based interactions with Amazon Alexa. By incorporating these modalities, our approach captures a more holistic representation of human behavior encompassing speech patterns, vocal characteristics, facial expressions, and linguistic structure, which may enable more accurate detection of cognitive decline. Different participants completed three conversational tasks designed to probe spontaneous speech, comprehension, and recall. We pursued three objectives: (1) evaluate which of the three Alexa tasks provides the most informative cognitive signals, (2) compare the effectiveness of unimodal models, and (3) assess the benefits of multimodal fusion models. Task-level analysis shows that the video model achieves its best performance on Command-Reading, with an accuracy of 0.65 (AUC 0.69), while the audio model performs best on Question-Answering, reaching an accuracy of 0.67 (AUC 0.73). In contrast, the multimodal model demonstrates consistently stable performance across tasks, achieving an accuracy of 0.76 (AUC 0.75) on Command-Reading and 0.76 (AUC 0.73) on Question-Answering. These findings demonstrate that multimodal behavioral cues captured through naturalistic interactions can reliably distinguish MCI from healthy aging, highlighting the limitations of traditional diagnostic paradigms and supporting the shift toward multimodal, real-world cognitive assessment frameworks.
Abstract (translated)
URL
https://arxiv.org/abs/2610.20906
PDF
-
SNOAR: Self-training with Noise filtering, Observation adding, and Remixing for Speech Denoising Without Clean References
2026-10-14 19:54:17Samuele Cornell, Matthew MaciejewskiAbstract
Unsupervised speech denoising methods such as MixIT and noisy-target training can be trained directly on noisy recordings, avoiding the sim-to-real gap of synthetic training data, but remain well below supervised performance. Teacher--student self-training, as in RemixIT, narrows this gap by remixing a teacher's estimates into new mixtures that supervise a student; however, since the student is trained only toward the teacher's estimates, teacher errors, in particular over-suppressed speech, are never recovered and propagate as the teacher is updated from the student. We propose SNOAR, a self-training framework that corrects the teacher's estimates before use: the noise leaked into its speech estimate is removed by differential noise filtering (DNF), and part of the noisy observation is added back in to restore suppressed speech and attenuate artifacts. We show that the two corrections depend on each other, as without DNF the loop collapses to the identity mapping. Compared to RemixIT, SNOAR is more stable along training and outperforms it in signal-based, perceptual, and word error rate metrics, reaching 97\% of the SI-SDR improvement of a fully-supervised topline on WHAM!; on real CHiME-3 recordings, it improves every metric of three different teachers, while RemixIT degrades their perceptual quality.
Abstract (translated)
URL
https://arxiv.org/abs/2610.20863
PDF
-
Zero-Shot Spoken Language Understanding with SLAM-ASR: What Do Monolingual ASR-Trained Linear Projectors Enable?
2026-10-14 19:28:20Edwin Arkel Rios, Sneh Sourabh, Jocelyn Zaruma, Girish Ewoorkar, Julian Mack, Bo-Cheng LaiAbstract
The SLAM-ASR framework connects the features of a frozen speech encoder as input to a frozen LLM decoder to enable speech transcription. In this work we ask whether this training-on speech transcription-provides the LLM with the capability to also perform spoken language understanding (SLU) tasks, by training monolingual projectors on 11 languages. When prompted to answer a yes/no question about the input speech, the mean instruction following rate (IFR) is 49.8%. When not following the instruction, these models default back to transcription. The transcription quality-measured by the character error rate (CER)-is not sufficient to predict IFR. Remarkably, when the model is instructed to answer the topic question as a single digit option, the mean IFR rises to 81.3%. Our models score a mean 23.2% topic balanced accuracy (BAcc) against 14.3% chance. However, the same decoder scores a mean 71.5% BAcc when using reference transcripts as text input. This indicates a failure of the projector to transfer information for the SLU tasks.
Abstract (translated)
URL
https://arxiv.org/abs/2610.20844
PDF
-
PEDBench: Can Speech Models Detect Phonetic Errors?
2026-10-14 19:15:48Zhonghao Shi, Zackary Rackauckas, Jasmine Yip, Savannah Halpern, Simon Campos, Lauren Brace, Emma Yang, Tiantian Feng, Xuan Shi, Shrikanth Narayanan, Teresa Satterfield, Ying XuAbstract
Phonetic error detection (PED) aims to identify phoneme-level substitutions, deletions, and insertions and is essential for educational and clinical speech assessment and intervention. Despite recent advances in speech models for phoneme recognition, their ability to reliably detect fine-grained phonetic errors remains unclear because of a lack of high-quality evaluation data. To address this gap, we introduce PEDBench, the first PED benchmark spanning a continuum from natural to controlled speech. It comprises real-world recordings of spoken errors made by children (N=334), scripted error productions by human actors (N=1,188), and synthetic error productions generated using phone-conditioned text-to-speech (N=1,083). All phonetic annotations were transcribed and verified by two phonetically trained annotators. Our evaluation of 13 state-of-the-art speech models shows that they fail to reliably detect phonetic errors, even under controlled conditions with acted and synthetic speech, where acoustic variability is reduced. The models also exhibit low precision, frequently misclassifying correctly pronounced phonemes as errors. Our speech-language models, fine-tuned on synthetic data, outperform commercial and open-source baselines on word-level PED, although a sim-to-real domain gap remains. We plan to publicly release our code and recipes, along with our acted and synthetic datasets, to support the development of more accurate PED systems.
Abstract (translated)
URL
https://arxiv.org/abs/2610.20830
PDF
-
Reference-Guided Turn-Taking Style Transfer for Full-Duplex Spoken Language Models
2026-10-14 18:12:58Yuxuan Chen, Yang Xiao, Jack Sun, Chak Sun, Brick Li, SamAbstract
Full-duplex spoken language models decide at every frame whether to take the floor, defining a conversational personality -- yet each set of weights encodes exactly one, which human speakers bracket rather than share. Existing controls mostly act from outside; none identifies the quantity that sets how much the model speaks. We identify it: a single late-layer residual-stream direction, the initiative vector, extracted from the model's own conversations without training or annotation. The same direction reads the decision, anticipating onsets up to 800 ms ahead, and shifts it: scaled during generation, it moves speech ratio from 6-12% to 42-61% (no steering: 21-30%), consistently across conversations, weights, and corpora. A short reference recording suffices to transfer a speaker's turn-taking style to the model, and the transfer is reference-specific: swapping in another speaker's reference worsens the match in 93-98% of cases, whereas the best uniform dose is indistinguishable from no steering. Turn-taking style is not frozen into the weights but a direction that can be read, shifted, and calibrated per person. Audio examples are available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2610.20763
PDF
-
EarlyCommit: Calibrated Word-by-Word Commitment for Robots Acting Before You Finish Speaking
Abstract
Human listeners can often identify the intended object several words before an instruction ends. Robots, however, are typically designed to wait until the speaker has finished before deciding how to act. We present EarlyCommit, an open-source tool that lets a robot commit to the intended object while the instruction is being spoken. After every recognized word, a fine-tuned text encoder scores each object in the scene. The robot acts when the calibrated probability of the best object reaches a threshold tau set by the user, so 1 - tau is the accepted risk of choosing the wrong object. Otherwise, it waits or asks. The tool includes a streaming speech front end, a scene and instruction generator with exact word-by-word ground truth, and scripts to train, calibrate, and evaluate. On instructions containing words never seen in training, EarlyCommit commits correctly 1.6 to 2 times as often as keyword matching, is right at least 92 percent of the time when it commits at tau = 0.9, and needs 31 to 35 ms per word on consumer GPUs.
Abstract (translated)
URL
https://arxiv.org/abs/2610.20531
PDF
-
Is Text All You Need? Text as a Universal Information Bottleneck for Speech LLMs
2026-10-14 16:52:10Ming-Hao Hsu, Yuxuan Hu, Shujie Liu, Jinyu Li, Yan Lu, Zhizheng WuAbstract
Large language models (LLMs) provide a powerful reasoning backbone for speech understanding, but integrating continuous acoustic signals into a pretrained LLM remains challenging. A central design question is how to tie speech representations to the LLM's text input space while preserving the continuous variation needed for different speech tasks. In this work, we propose Convex Gate (C-Gate), a speech-LLM bridge that represents each frame as a convex combination of frozen token embeddings. This architectural convex-hull constraint fixes the output codebook while allowing continuous mixtures learned from task supervision. With a Whisper-large-v3 encoder, a partially adapted Qwen2.5-7B-Instruct decoder, and joint training on speech recognition, emotion recognition, and spoken reasoning, C-Gate performs accurate transcription and emotion recognition, our representative semantic and acoustic tasks, through one shared decoder. On text-only benchmarks, the adapted decoder loses instruction-following and MMLU accuracy and matches or exceeds the original model on GSM8K and HellaSwag. We further design analyses and interventions on the trained model to examine where the bridge sequence carries task information and how the decoder reads it. These findings show how fixed text embeddings can support both semantic and acoustic readouts through continuous speech-conditioned mixtures.
Abstract (translated)
URL
https://arxiv.org/abs/2606.09366
PDF
