Whisper-CD: Accurate Long-Form Speech Recognition using Multi-Negative Contrastive Decoding

• URL: http://arxiv.org/abs/2603.06193v1
• Date: Fri, 06 Mar 2026 12:04:12 GMT
• Title: Whisper-CD: Accurate Long-Form Speech Recognition using Multi-Negative Contrastive Decoding
• Abstract summary: We propose Whisper-CD, a training-free contrastive decoding framework that contrasts clean-audio logits against negative logits computed from three acoustically motivated perturbations.<n>We aggregate these negatives via the log-sum-exp operator, building a unified multi-negative objective for token-by-token decoding.<n>Whisper-CD reduces WER by up to 24.3pp on CORAAL and shows 48% faster token generation throughput than beam search.
• Score: 10.438946760176787
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Long-form speech recognition with large encoder-decoder models such as Whisper often exhibit hallucinations, repetition loops, and content omissions. These errors can accumulate and be further amplified when the previous segment's transcription is used as decoding context. We propose Whisper-CD, a training-free contrastive decoding framework that contrasts clean-audio logits against negative logits computed from three acoustically motivated perturbations: Gaussian noise injection, silence signal, and audio temporal shift. We aggregate these negatives via the log-sum-exp operator, building a unified multi-negative objective for token-by-token decoding. Across five English long-form benchmarks, Whisper-CD reduces WER by up to 24.3pp on CORAAL and shows 48% faster token generation throughput than beam search. Because Whisper-CD operates purely at inference time, it can be applied as a drop-in replacement to already-deployed Whisper systems without retraining.

Related papers

• Stride-k Subsampling: Train-Free Audio Token Reduction for Whisper [4.214795904343116]
  We propose stride-k subsampling, a deterministic indexing operation that retains every k-th token after the convolutional stem or encoder transformer.<n>Applying stride-2 at both positions cuts audio tokens by 75% and total GFLOPs by 52-58%, with small WER costs on most ASR benchmarks and larger costs on harder ones.<n>The same configuration extends to three Whisper-based SpeechLMs, yielding modest accuracy drops on stronger baselines and larger drops on weaker ones, while reducing end-to-end latency by 19.6-27.4%.
  arXiv  Detail & Related papers  (2026-08-31T15:04:29Z)
• Beyond U-Net: A Latent-Representation-Aligned Skip-Free Backbone for Flow-Matching Speech Enhancement [1.8907108368038215]
  We propose a skip-free encoder-decoder backbone for flow-matching speech enhancement, guided by Latent Representation Alignment (LRA)<n>Experiments on WSJ0-CHiME3 and VoiceBank-DEMAND show improved PESQ and perceptual quality, especially on VoiceBank-DEMAND, using only five function evaluations.
  arXiv  Detail & Related papers  (2026-06-23T16:09:31Z)
• AudioMosaic: Contrastive Masked Audio Representation Learning [53.52371029884106]
  We introduce textbfAudioMosaic, a contrastive learning-based audio encoder for general audio understanding.<n>AudioMosaic constructs positive pairs by applying structured time-frequency masking to spectrogram patches.<n>Experiments show that AudioMosaic achieves state-of-the-art performance on several standard audio benchmarks.
  arXiv  Detail & Related papers  (2026-05-14T00:56:51Z)
• Temporal Contrastive Decoding: A Training-Free Method for Large Audio-Language Models [56.91801348360746]
  Large audio-language models (LALMs) generalize across speech, sound, and music.<n> unified decoders can exhibit a emphtemporal smoothing bias<n>We propose emphTemporal Contrastive Decoding (TCD), a training-free decoding method for unified LALMs.
  arXiv  Detail & Related papers  (2026-04-16T02:30:41Z)
• TASTE-Streaming: Towards Streamable Text-Aligned Speech Tokenization and Embedding for Spoken Language Modeling [52.519745466772825]
  Text-speech joint spoken language modeling aims at natural and intelligent speech-based interactions.<n>TASTE produces speech tokens that align in lengths with their textual counterparts.<n>We propose TASTE-S, a streamable extension of TASTE suitable for real-time usage.
  arXiv  Detail & Related papers  (2026-03-12T18:13:48Z)
• CrisperWhisper: Accurate Timestamps on Verbatim Speech Transcriptions [0.5120567378386615]
  We fine-tune the model to produce more verbatim speech transcriptions. We employ several techniques to increase robustness against multiple speakers and background noise.
  arXiv  Detail & Related papers  (2024-08-29T14:52:42Z)
• Simul-Whisper: Attention-Guided Streaming Whisper with Truncation Detection [14.989929439744172]
  We introduce Simul-Whisper, which uses the time alignment embedded in Whisper's cross-attention to guide auto-regressive decoding. Experiments on multiple languages and Whisper architectures show that Simul-Whisper achieves an average absolute word error rate of only 1.46% at a chunk size of 1 second.
  arXiv  Detail & Related papers  (2024-06-14T14:07:26Z)
• A Non-autoregressive Generation Framework for End-to-End Simultaneous Speech-to-Speech Translation [48.84039953531355]
  We propose a novel non-autoregressive generation framework for simultaneous speech translation (NAST-S2X) NAST-S2X integrates speech-to-text and speech-to-speech tasks into a unified end-to-end framework. It achieves high-quality simultaneous interpretation within a delay of less than 3 seconds and provides a 28 times decoding speedup in offline generation.
  arXiv  Detail & Related papers  (2024-06-11T04:25:48Z)
• Autoregressive Diffusion Transformer for Text-to-Speech Synthesis [39.32761051774537]
  We propose encoding audio as vector sequences in continuous space $mathbb Rd$ and autoregressively generating these sequences. High-bitrate continuous speech representation enables almost flawless reconstruction, allowing our model to achieve nearly perfect speech editing.
  arXiv  Detail & Related papers  (2024-06-08T18:57:13Z)
• Diffsound: Discrete Diffusion Model for Text-to-sound Generation [78.4128796899781]
  We propose a novel text-to-sound generation framework that consists of a text encoder, a Vector Quantized Variational Autoencoder (VQ-VAE), a decoder, and a vocoder. The framework first uses the decoder to transfer the text features extracted from the text encoder to a mel-spectrogram with the help of VQ-VAE, and then the vocoder is used to transform the generated mel-spectrogram into a waveform.
  arXiv  Detail & Related papers  (2022-07-20T15:41:47Z)
• Pre-Training Transformer Decoder for End-to-End ASR Model with Unpaired Speech Data [145.95460945321253]
  We introduce two pre-training tasks for the encoder-decoder network using acoustic units, i.e., pseudo codes. The proposed Speech2C can relatively reduce the word error rate (WER) by 19.2% over the method without decoder pre-training.
  arXiv  Detail & Related papers  (2022-03-31T15:33:56Z)
• Attentional Speech Recognition Models Misbehave on Out-of-domain Utterances [16.639133822656458]
  We decode audio from the British National Corpus with an attentional encoder-decoder model trained solely on the LibriSpeech corpus. We observe that there are many 5-second recordings that produce more than 500 characters of decoding output. A frame-synchronous hybrid (DNN-HMM) model trained on the same data does not produce these unusually long transcripts.
  arXiv  Detail & Related papers  (2020-02-12T18:53:56Z)
• Multi-task self-supervised learning for Robust Speech Recognition [75.11748484288229]
  This paper proposes PASE+, an improved version of PASE for robust speech recognition in noisy and reverberant environments. We employ an online speech distortion module, that contaminates the input signals with a variety of random disturbances. We then propose a revised encoder that better learns short- and long-term speech dynamics with an efficient combination of recurrent and convolutional networks.
  arXiv  Detail & Related papers  (2020-01-25T00:24:45Z)

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.