SITA: Learning Speaker-Invariant and Tone-Aware Speech Representations for Low-Resource Tonal Languages

• URL: http://arxiv.org/abs/2601.09050v1
• Date: Wed, 14 Jan 2026 00:42:27 GMT
• Title: SITA: Learning Speaker-Invariant and Tone-Aware Speech Representations for Low-Resource Tonal Languages
• Authors: Tianyi Xu, Xuan Ouyang, Binwei Yao, Shoua Xiong, Sara Misurelli, Maichou Lor, Junjie Hu,
• Abstract summary: SITA is a lightweight adaptation recipe that enforces Speaker-Invariance and Tone-Awareness for pretrained wav2vec-style encoders.<n>We evaluate primarily on Hmong, a highly tonal and severely under-the-shelf multilingual encoders fail to represent tone effectively.
• Score: 11.655315357810371
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Tonal low-resource languages are widely spoken yet remain underserved by modern speech technology. A key challenge is learning representations that are robust to nuisance variation such as gender while remaining tone-aware for different lexical meanings. To address this, we propose SITA, a lightweight adaptation recipe that enforces Speaker-Invariance and Tone-Awareness for pretrained wav2vec-style encoders. SITA uses staged multi-objective training: (i) a cross-gender contrastive objective encourages lexical consistency across speakers, while a tone-repulsive loss prevents tone collapse by explicitly separating same-word different-tone realizations; and (ii) an auxiliary Connectionist Temporal Classification (CTC)-based ASR objective with distillation stabilizes recognition-relevant structure. We evaluate primarily on Hmong, a highly tonal and severely under-resourced language where off-the-shelf multilingual encoders fail to represent tone effectively. On a curated Hmong word corpus, SITA improves cross-gender lexical retrieval accuracy, while maintaining usable ASR accuracy relative to an ASR-adapted XLS-R teacher. We further observe similar gains when transferring the same recipe to Mandarin, suggesting SITA is a general, plug-in approach for adapting multilingual speech encoders to tonal languages.

Related papers

• Towards Unsupervised Speech Recognition at the Syllable-Level [95.54031547995874]
  We introduce a syllable-level UASR framework based on masked language modeling.<n>We generalize effectively to Mandarin, a language that has remained particularly difficult for prior methods.
  arXiv  Detail & Related papers  (2025-10-04T02:56:33Z)
• PART: Progressive Alignment Representation Training for Multilingual Speech-To-Text with LLMs [58.2469845374385]
  We introduce Progressive Alignment Representation Training (PART)<n>PART is a multi-stage and multi-task framework that separates within-language from cross-language alignment.<n>Experiments on CommonVoice 15, Fleurs, Wenetspeech, and CoVoST2 show that PART surpasses conventional approaches.
  arXiv  Detail & Related papers  (2025-09-24T03:54:14Z)
• LASPA: Language Agnostic Speaker Disentanglement with Prefix-Tuned Cross-Attention [2.199918533021483]
  The overlap between vocal traits such as accent, vocal anatomy, and a language's phonetic structure complicates separating linguistic and speaker information.<n>Disentangling these components can significantly improve speaker recognition accuracy.<n>We propose a novel disentanglement learning strategy that integrates joint learning through prefix-tuned cross-attention.
  arXiv  Detail & Related papers  (2025-06-02T10:59:31Z)
• Whisper Speaker Identification: Leveraging Pre-Trained Multilingual Transformers for Robust Speaker Embeddings [0.0]
  We propose WSI (Whisper Speaker Identification), a framework that repurposes the Whisper automatic speech recognition model pre trained on extensive multilingual data.<n>By capitalizing on Whisper language-agnostic acoustic representations, our approach effectively distinguishes speakers across diverse languages.
  arXiv  Detail & Related papers  (2025-03-13T15:11:28Z)
• Leveraging Language ID to Calculate Intermediate CTC Loss for Enhanced Code-Switching Speech Recognition [5.3545957730615905]
  We introduce language identification information into the middle layer of the ASR model's encoder. We aim to generate acoustic features that imply language distinctions in a more implicit way, reducing the model's confusion when dealing with language switching.
  arXiv  Detail & Related papers  (2023-12-15T07:46:35Z)
• Simple yet Effective Code-Switching Language Identification with Multitask Pre-Training and Transfer Learning [0.7242530499990028]
  Code-switching is the linguistics phenomenon where in casual settings, multilingual speakers mix words from different languages in one utterance. We propose two novel approaches toward improving language identification accuracy on an English-Mandarin child-directed speech dataset. Our best model achieves a balanced accuracy of 0.781 on a real English-Mandarin code-switching child-directed speech corpus and outperforms the previous baseline by 55.3%.
  arXiv  Detail & Related papers  (2023-05-31T11:43:16Z)
• The Interpreter Understands Your Meaning: End-to-end Spoken Language Understanding Aided by Speech Translation [13.352795145385645]
  Speech translation (ST) is a good means of pretraining speech models for end-to-end spoken language understanding. We show that our models reach higher performance over baselines on monolingual and multilingual intent classification. We also create new benchmark datasets for speech summarization and low-resource/zero-shot transfer from English to French or Spanish.
  arXiv  Detail & Related papers  (2023-05-16T17:53:03Z)
• LAMASSU: Streaming Language-Agnostic Multilingual Speech Recognition and Translation Using Neural Transducers [71.76680102779765]
  Automatic speech recognition (ASR) and speech translation (ST) can both use neural transducers as the model structure. We propose LAMASSU, a streaming language-agnostic multilingual speech recognition and translation model using neural transducers.
  arXiv  Detail & Related papers  (2022-11-05T04:03:55Z)
• 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)
• Mandarin-English Code-switching Speech Recognition with Self-supervised Speech Representation Models [55.82292352607321]
  Code-switching (CS) is common in daily conversations where more than one language is used within a sentence. This paper uses the recently successful self-supervised learning (SSL) methods to leverage many unlabeled speech data without CS.
  arXiv  Detail & Related papers  (2021-10-07T14:43:35Z)
• How Phonotactics Affect Multilingual and Zero-shot ASR Performance [74.70048598292583]
  A Transformer encoder-decoder model has been shown to leverage multilingual data well in IPA transcriptions of languages presented during training. We replace the encoder-decoder with a hybrid ASR system consisting of a separate AM and LM. We show that the gain from modeling crosslingual phonotactics is limited, and imposing a too strong model can hurt the zero-shot transfer.
  arXiv  Detail & Related papers  (2020-10-22T23:07:24Z)
• Learning not to Discriminate: Task Agnostic Learning for Improving Monolingual and Code-switched Speech Recognition [12.354292498112347]
  We present further improvements over our previous work by using domain adversarial learning to train task models. Our proposed technique leads to reductions in Word Error Rates (WER) in monolingual and code-switched test sets across three language pairs.
  arXiv  Detail & Related papers  (2020-06-09T13:45:30Z)
• That Sounds Familiar: an Analysis of Phonetic Representations Transfer Across Languages [72.9927937955371]
  We use the resources existing in other languages to train a multilingual automatic speech recognition model. We observe significant improvements across all languages in the multilingual setting, and stark degradation in the crosslingual setting. Our analysis uncovered that even the phones that are unique to a single language can benefit greatly from adding training data from other languages.
  arXiv  Detail & Related papers  (2020-05-16T22:28:09Z)

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.