Recent advances in deep neural networks have achieved unprecedented success in visual speech recognition. However, there remains substantial disparity between current methods and their deployment in resource-constrained devices. In this work, we explore different magnitude-based pruning techniques to generate a lightweight model that achieves higher performance than its dense model equivalent, especially under the presence of visual noise. Our sparse models achieve state-of-the-art results at 10% sparsity on the LRS3 dataset and outperform the dense equivalent up to 70% sparsity. We evaluate our 50% sparse model on 7 different visual noise types and achieve an overall absolute improvement of more than 2% WER compared to the dense equivalent. Our results confirm that sparse networks are more resistant to noise than dense networks.
Research in multilingual speech-to-text translation is topical. Having a single model that supports multiple translation tasks is desirable. The goal of this work it to improve cross-lingual transfer learning in multilingual speech-to-text translation via semantic knowledge distillation. We show that by initializing the encoder of the encoder-decoder sequence-to-sequence translation model with SAMU-XLS-R, a multilingual speech transformer encoder trained using multi-modal (speech-text) semantic knowledge distillation, we achieve significantly better cross-lingual task knowledge transfer than the baseline XLS-R, a multilingual speech transformer encoder trained via self-supervised learning. We demonstrate the effectiveness of our approach on two popular datasets, namely, CoVoST-2 and Europarl. On the 21 translation tasks of the CoVoST-2 benchmark, we achieve an average improvement of 12.8 BLEU points over the baselines. In the zero-shot translation scenario, we achieve an average gain of 18.8 and 11.9 average BLEU points on unseen medium and low-resource languages. We make similar observations on Europarl speech translation benchmark.
Audio deepfake detection (ADD) is the task of detecting spoofing attacks generated by text-to-speech or voice conversion systems. Spoofing evidence, which helps to distinguish between spoofed and bona-fide utterances, might exist either locally or globally in the input features. To capture these, the Conformer, which consists of Transformers and CNN, possesses a suitable structure. However, since the Conformer was designed for sequence-to-sequence tasks, its direct application to ADD tasks may be sub-optimal. To tackle this limitation, we propose HM-Conformer by adopting two components: (1) Hierarchical pooling method progressively reducing the sequence length to eliminate duplicated information (2) Multi-level classification token aggregation method utilizing classification tokens to gather information from different blocks. Owing to these components, HM-Conformer can efficiently detect spoofing evidence by processing various sequence lengths and aggregating them. In experimental results on the ASVspoof 2021 Deepfake dataset, HM-Conformer achieved a 15.71% EER, showing competitive performance compared to recent systems.
High-quality human transcription is essential for training and improving Automatic Speech Recognition (ASR) models. Recent study~\cite{libricrowd} has found that every 1% worse transcription Word Error Rate (WER) increases approximately 2% ASR WER by using the transcriptions to train ASR models. Transcription errors are inevitable for even highly-trained annotators. However, few studies have explored human transcription correction. Error correction methods for other problems, such as ASR error correction and grammatical error correction, do not perform sufficiently for this problem. Therefore, we propose HTEC for Human Transcription Error Correction. HTEC consists of two stages: Trans-Checker, an error detection model that predicts and masks erroneous words, and Trans-Filler, a sequence-to-sequence generative model that fills masked positions. We propose a holistic list of correction operations, including four novel operations handling deletion errors. We further propose a variant of embeddings that incorporates phoneme information into the input of the transformer. HTEC outperforms other methods by a large margin and surpasses human annotators by 2.2% to 4.5% in WER. Finally, we deployed HTEC to assist human annotators and showed HTEC is particularly effective as a co-pilot, which improves transcription quality by 15.1% without sacrificing transcription velocity.
Large language models (LLMs) have gained considerable attention for Artificial Intelligence Generated Content (AIGC), particularly with the emergence of ChatGPT. However, the direct adaptation of continuous speech to LLMs that process discrete tokens remains an unsolved challenge, hindering the application of LLMs for speech generation. The advanced speech LMs are in the corner, as that speech signals encapsulate a wealth of information, including speaker and emotion, beyond textual data alone. Prompt tuning has demonstrated notable gains in parameter efficiency and competitive performance on some speech classification tasks. However, the extent to which prompts can effectively elicit generation tasks from speech LMs remains an open question. In this paper, we present pioneering research that explores the application of prompt tuning to stimulate speech LMs for various generation tasks, within a unified framework called SpeechGen, with around 10M trainable parameters. The proposed unified framework holds great promise for efficiency and effectiveness, particularly with the imminent arrival of advanced speech LMs, which will significantly enhance the capabilities of the framework. The code and demos of SpeechGen will be available on the project website: \url{https://ga642381.github.io/SpeechPrompt/speechgen}
Self-supervised learning (SSL) is a popular research topic in speech processing. Successful SSL speech models must generalize well. SUPERB was proposed to evaluate the ability of SSL speech models across many speech tasks. However, due to the diversity of tasks, the evaluation process requires huge computational costs. We present MiniSUPERB, a lightweight benchmark that efficiently evaluates SSL speech models with comparable results to SUPERB while greatly reducing the computational cost. We select representative tasks and sample datasets and extract model representation offline, achieving 0.954 and 0.982 Spearman's rank correlation with SUPERB Paper and SUPERB Challenge, respectively. In the meanwhile, the computational cost is reduced by 97% in regard to MACs (number of Multiply-ACcumulate operations) in the tasks we choose. To the best of our knowledge, this is the first study to examine not only the computational cost of a model itself but the cost of evaluating it on a benchmark.
Rich sources of variability in natural speech present significant challenges to current data intensive speech recognition technologies. To model both speaker and environment level diversity, this paper proposes a novel Bayesian factorised speaker-environment adaptive training and test time adaptation approach for Conformer ASR models. Speaker and environment level characteristics are separately modeled using compact hidden output transforms, which are then linearly or hierarchically combined to represent any speaker-environment combination. Bayesian learning is further utilized to model the adaptation parameter uncertainty. Experiments on the 300-hr WHAM noise corrupted Switchboard data suggest that factorised adaptation consistently outperforms the baseline and speaker label only adapted Conformers by up to 3.1% absolute (10.4% relative) word error rate reductions. Further analysis shows the proposed method offers potential for rapid adaption to unseen speaker-environment conditions.
Despite recent advancements in speech recognition, there are still difficulties in accurately transcribing conversational and emotional speech in noisy and reverberant acoustic environments. This poses a particular challenge in the search and rescue (SAR) domain, where transcribing conversations among rescue team members is crucial to support real-time decision-making. The scarcity of speech data and associated background noise in SAR scenarios make it difficult to deploy robust speech recognition systems. To address this issue, we have created and made publicly available a German speech dataset called RescueSpeech. This dataset includes real speech recordings from simulated rescue exercises. Additionally, we have released competitive training recipes and pre-trained models. Our study indicates that the current level of performance achieved by state-of-the-art methods is still far from being acceptable.
Recognition of personalized content remains a challenge in end-to-end speech recognition. We explore three novel approaches that use personalized content in a neural rescoring step to improve recognition: gazetteers, prompting, and a cross-attention based encoder-decoder model. We use internal de-identified en-US data from interactions with a virtual voice assistant supplemented with personalized named entities to compare these approaches. On a test set with personalized named entities, we show that each of these approaches improves word error rate by over 10%, against a neural rescoring baseline. We also show that on this test set, natural language prompts can improve word error rate by 7% without any training and with a marginal loss in generalization. Overall, gazetteers were found to perform the best with a 10% improvement in word error rate (WER), while also improving WER on a general test set by 1%.
In this paper, we aim to address the problem of channel robustness in speech countermeasure (CM) systems, which are used to distinguish synthetic speech from human natural speech. On the basis of two hypotheses, we suggest an approach for perturbing phase information during the training of time-domain CM systems. Communication networks often employ lossy compression codec that encodes only magnitude information, therefore heavily altering phase information. Also, state-of-the-art CM systems rely on phase information to identify spoofed speech. Thus, we believe the information loss in the phase domain induced by lossy compression codec degrades the performance of the unseen channel. We first establish the dependence of time-domain CM systems on phase information by perturbing phase in evaluation, showing strong degradation. Then, we demonstrated that perturbing phase during training leads to a significant performance improvement, whereas perturbing magnitude leads to further degradation.