Bridging What the Model Thinks and How It Speaks: Self-Aware Speech Language Models for Expressive Speech Generation

• URL: http://arxiv.org/abs/2604.11424v1
• Date: Mon, 13 Apr 2026 13:06:09 GMT
• Title: Bridging What the Model Thinks and How It Speaks: Self-Aware Speech Language Models for Expressive Speech Generation
• Abstract summary: Speech Language Models (SLMs) exhibit strong semantic understanding, yet their generated speech often sounds flat and fails to convey expressive intent, undermining user engagement.<n>We term this mismatch the semantic understanding-acoustic realization gap.<n>We propose SA-SLM (Self-Aware Speech Language Model), built on the principle that the model should be aware of what it thinks during generation and how it speaks during training.
• Score: 34.357436199397604
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Speech Language Models (SLMs) exhibit strong semantic understanding, yet their generated speech often sounds flat and fails to convey expressive intent, undermining user engagement. We term this mismatch the semantic understanding-acoustic realization gap. We attribute this gap to two key deficiencies: (1) intent transmission failure, where SLMs fail to provide the stable utterance-level intent needed for expressive delivery; and (2) realization-unaware training, where no feedback signal verifies whether acoustic outputs faithfully reflect intended expression. To address these issues, we propose SA-SLM (Self-Aware Speech Language Model), built on the principle that the model should be aware of what it thinks during generation and how it speaks during training. SA-SLM addresses this gap through two core contributions: (1) Intent-Aware Bridging, which uses a Variational Information Bottleneck (VIB) objective to translate the model's internal semantics into temporally smooth expressive intent, making speech generation aware of what the model intends to express; and (2) Realization-Aware Alignment, which repurposes the model as its own critic to verify and align acoustic realization with intended expressive intent via rubric-based feedback. Trained on only 800 hours of expressive speech data, our 3B parameter SA-SLM surpasses all open-source baselines and comes within 0.08 points of GPT-4o-Audio in overall expressiveness on the EchoMind benchmark.

Related papers

• DiffuSpeech: Silent Thought, Spoken Answer via Unified Speech-Text Diffusion [23.01044837428522]
  We introduce textbfSilent Thought, Spoken Answer'' -- a paradigm where speech LLMs generate internal text reasoning alongside spoken responses.<n>We present method, the first diffusion-based speech-text language model supporting both understanding and generation.<n>Experiments show method achieves state-of-the-art speech-to-speech QA accuracy, outperforming the best baseline by up to 9 points.
  arXiv  Detail & Related papers  (2026-01-30T12:08:33Z)
• Optimizing Speech Language Models for Acoustic Consistency [2.5864269455844484]
  We train three models: a 0.7B speech-only model, a 1.0B speech-only model, and a 1.0B interleaved model with both text and speech.<n>Our approach initializes speech tokens with self-supervised features, applies a light alignment loss, and trains with thinning and auxiliary objectives.
  arXiv  Detail & Related papers  (2025-09-30T13:59:52Z)
• S2S-Arena, Evaluating Speech2Speech Protocols on Instruction Following with Paralinguistic Information [47.950757976473035]
  We introduce S2S-Arena, a novel arena-style S2S benchmark that evaluates instruction-following capabilities with paralinguistic information.<n>In addition to the superior performance of GPT-4o, the speech model of cascaded ASR, LLM, and TTS outperforms the jointly trained model after text-speech alignment in speech2speech protocols.
  arXiv  Detail & Related papers  (2025-03-07T02:07:00Z)
• Integrating Self-supervised Speech Model with Pseudo Word-level Targets from Visually-grounded Speech Model [57.78191634042409]
  We propose Pseudo-Word HuBERT (PW-HuBERT), a framework that integrates pseudo word-level targets into the training process. Our experimental results on four spoken language understanding (SLU) benchmarks suggest the superiority of our model in capturing semantic information.
  arXiv  Detail & Related papers  (2024-02-08T16:55:21Z)
• Disentangling Voice and Content with Self-Supervision for Speaker Recognition [57.446013973449645]
  This paper proposes a disentanglement framework that simultaneously models speaker traits and content variability in speech. It is validated with experiments conducted on the VoxCeleb and SITW datasets with 9.56% and 8.24% average reductions in EER and minDCF.
  arXiv  Detail & Related papers  (2023-10-02T12:02:07Z)
• Improving the Gap in Visual Speech Recognition Between Normal and Silent Speech Based on Metric Learning [11.50011780498048]
  This paper presents a novel metric learning approach to address the performance gap between normal and silent speech in visual speech recognition (VSR) We propose to leverage the shared literal content between normal and silent speech and present a metric learning approach based on visemes. Our evaluation demonstrates that our method improves the accuracy of silent VSR, even when limited training data is available.
  arXiv  Detail & Related papers  (2023-05-23T16:20:46Z)
• Towards Language Modelling in the Speech Domain Using Sub-word Linguistic Units [56.52704348773307]
  We propose a novel LSTM-based generative speech LM based on linguistic units including syllables and phonemes. With a limited dataset, orders of magnitude smaller than that required by contemporary generative models, our model closely approximates babbling speech. We show the effect of training with auxiliary text LMs, multitask learning objectives, and auxiliary articulatory features.
  arXiv  Detail & Related papers  (2021-10-31T22:48:30Z)
• Wav2vec-Switch: Contrastive Learning from Original-noisy Speech Pairs for Robust Speech Recognition [52.71604809100364]
  We propose wav2vec-Switch, a method to encode noise robustness into contextualized representations of speech. Specifically, we feed original-noisy speech pairs simultaneously into the wav2vec 2.0 network. In addition to the existing contrastive learning task, we switch the quantized representations of the original and noisy speech as additional prediction targets.
  arXiv  Detail & Related papers  (2021-10-11T00:08:48Z)
• Pre-training for Spoken Language Understanding with Joint Textual and Phonetic Representation Learning [4.327558819000435]
  We propose a novel joint textual-phonetic pre-training approach for learning spoken language representations. Experimental results on spoken language understanding benchmarks, Fluent Speech Commands and SNIPS, show that the proposed approach significantly outperforms strong baseline models.
  arXiv  Detail & Related papers  (2021-04-21T05:19:13Z)
• SPLAT: Speech-Language Joint Pre-Training for Spoken Language Understanding [61.02342238771685]
  Spoken language understanding requires a model to analyze input acoustic signal to understand its linguistic content and make predictions. Various pre-training methods have been proposed to learn rich representations from large-scale unannotated speech and text. We propose a novel semi-supervised learning framework, SPLAT, to jointly pre-train the speech and language modules.
  arXiv  Detail & Related papers  (2020-10-05T19:29:49Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.