arXiv subjects

Shinji Watanabe

Publications and source records attributed to Shinji Watanabe.

At least 181 records · Page 10Linked to original sources

arXiv2024

Decoder-only Architecture for Streaming End-to-end Speech Recognition

Decoder-only language models (LMs) have been successfully adopted for speech-processing tasks including automatic speech recognition (ASR). The LMs have ample expressiveness and perform efficiently. This efficiency is a suitable characteristic for streaming applications of ASR. In this work, we propose to use a decoder-only architecture for blockwise streaming ASR. In our approach, speech features are compressed using CTC output and context embedding using blockwise speech subnetwork, and are sequentially provided as prompts to the decoder. The decoder estimates the output tokens promptly at each block. To this end, we also propose a novel training scheme using random-length prefix prompts to make the model robust to the truncated prompts caused by blockwise processing. An experimental comparison shows that our proposed decoder-only streaming ASR achieves 8% relative word error rate reduction in the LibriSpeech test-other set while being twice as fast as the baseline model.

eess.AS↗

arXiv2024

SynesLM: A Unified Approach for Audio-visual Speech Recognition and Translation via Language Model and Synthetic Data

In this work, we present SynesLM, an unified model which can perform three multimodal language understanding tasks: audio-visual automatic speech recognition(AV-ASR) and visual-aided speech/machine translation(VST/VMT). Unlike previous research that focused on lip motion as visual cues for speech signals, our work explores more general visual information within entire frames, such as objects and actions. Additionally, we use synthetic image data to enhance the correlation between image and speech data. We benchmark SynesLM against the How2 dataset, demonstrating performance on par with state-of-the-art (SOTA) models dedicated to AV-ASR while maintaining our multitasking framework. Remarkably, for zero-shot AV-ASR, SynesLM achieved SOTA performance by lowering the Word Error Rate (WER) from 43.4% to 39.4% on the VisSpeech Dataset. Furthermore, our results in VST and VMT outperform the previous results, improving the BLEU score to 43.5 from 37.2 for VST, and to 54.8 from 54.4 for VMT.

eess.AS↗

arXiv2024

Multi-Convformer: Extending Conformer with Multiple Convolution Kernels

Convolutions have become essential in state-of-the-art end-to-end Automatic Speech Recognition~(ASR) systems due to their efficient modelling of local context. Notably, its use in Conformers has led to superior performance compared to vanilla Transformer-based ASR systems. While components other than the convolution module in the Conformer have been reexamined, altering the convolution module itself has been far less explored. Towards this, we introduce Multi-Convformer that uses multiple convolution kernels within the convolution module of the Conformer in conjunction with gating. This helps in improved modeling of local dependencies at varying granularities. Our model rivals existing Conformer variants such as CgMLP and E-Branchformer in performance, while being more parameter efficient. We empirically compare our approach with Conformer and its variants across four different datasets and three different modelling paradigms and show up to 8% relative word error rate~(WER) improvements.

cs.CL↗

arXiv2024

The CHiME-8 DASR Challenge for Generalizable and Array Agnostic Distant Automatic Speech Recognition and Diarization

This paper presents the CHiME-8 DASR challenge which carries on from the previous edition CHiME-7 DASR (C7DASR) and the past CHiME-6 challenge. It focuses on joint multi-channel distant speech recognition (DASR) and diarization with one or more, possibly heterogeneous, devices. The main goal is to spur research towards meeting transcription approaches that can generalize across arbitrary number of speakers, diverse settings (formal vs. informal conversations), meeting duration, wide-variety of acoustic scenarios and different recording configurations. Novelties with respect to C7DASR include: i) the addition of NOTSOFAR-1, an additional office/corporate meeting scenario, ii) a manually corrected Mixer 6 development set, iii) a new track in which we allow the use of large-language models (LLM) iv) a jury award mechanism to encourage participants to explore also more practical and innovative solutions. To lower the entry barrier for participants, we provide a standalone toolkit for downloading and preparing such datasets as well as performing text normalization and scoring their submissions. Furthermore, this year we also provide two baseline systems, one directly inherited from C7DASR and based on ESPnet and another one developed on NeMo and based on NeMo team submission in last year C7DASR. Baseline system results suggest that the addition of the NOTSOFAR-1 scenario significantly increases the task's difficulty due to its high number of speakers and very short duration.

eess.AS↗

arXiv2024

Less Peaky and More Accurate CTC Forced Alignment by Label Priors

Connectionist temporal classification (CTC) models are known to have peaky output distributions. Such behavior is not a problem for automatic speech recognition (ASR), but it can cause inaccurate forced alignments (FA), especially at finer granularity, e.g., phoneme level. This paper aims at alleviating the peaky behavior for CTC and improve its suitability for forced alignment generation, by leveraging label priors, so that the scores of alignment paths containing fewer blanks are boosted and maximized during training. As a result, our CTC model produces less peaky posteriors and is able to more accurately predict the offset of the tokens besides their onset. It outperforms the standard CTC model and a heuristics-based approach for obtaining CTC's token offset timestamps by 12-40% in phoneme and word boundary errors (PBE and WBE) measured on the Buckeye and TIMIT data. Compared with the most widely used FA toolkit Montreal Forced Aligner (MFA), our method performs similarly on PBE/WBE on Buckeye, yet falls behind MFA on TIMIT. Nevertheless, our method has a much simpler training pipeline and better runtime efficiency. Our training recipe and pretrained model are released in TorchAudio.

eess.AS↗

arXiv2024

Towards Robust Speech Representation Learning for Thousands of Languages

Self-supervised learning (SSL) has helped extend speech technologies to more languages by reducing the need for labeled data. However, models are still far from supporting the world's 7000+ languages. We propose XEUS, a Cross-lingual Encoder for Universal Speech, trained on over 1 million hours of data across 4057 languages, extending the language coverage of SSL models 4-fold. We combine 1 million hours of speech from existing publicly accessible corpora with a newly created corpus of 7400+ hours from 4057 languages, which will be publicly released. To handle the diverse conditions of multilingual speech data, we augment the typical SSL masked prediction approach with a novel dereverberation objective, increasing robustness. We evaluate XEUS on several benchmarks, and show that it consistently outperforms or achieves comparable results to state-of-the-art (SOTA) SSL models across a variety of tasks. XEUS sets a new SOTA on the ML-SUPERB benchmark: it outperforms MMS 1B and w2v-BERT 2.0 v2 by 0.8% and 4.4% respectively, despite having less parameters or pre-training data. Checkpoints, code, and data are found in https://www.wavlab.org/activities/2024/xeus/.

cs.CL↗

arXiv2024

Contextualized End-to-end Automatic Speech Recognition with Intermediate Biasing Loss

Contextualized end-to-end automatic speech recognition has been an active research area, with recent efforts focusing on the implicit learning of contextual phrases based on the final loss objective. However, these approaches ignore the useful contextual knowledge encoded in the intermediate layers. We hypothesize that employing explicit biasing loss as an auxiliary task in the encoder intermediate layers may better align text tokens or audio frames with the desired objectives. Our proposed intermediate biasing loss brings more regularization and contextualization to the network. Our method outperforms a conventional contextual biasing baseline on the LibriSpeech corpus, achieving a relative improvement of 22.5% in biased word error rate (B-WER) and up to 44% compared to the non-contextual baseline with a biasing list size of 100. Moreover, employing RNN-transducer-driven joint decoding further reduces the unbiased word error rate (U-WER), resulting in a more robust network.

eess.AS↗

arXiv2024

Diffusion-based Generative Modeling with Discriminative Guidance for Streamable Speech Enhancement

Diffusion-based generative models (DGMs) have recently attracted attention in speech enhancement research (SE) as previous works showed a remarkable generalization capability. However, DGMs are also computationally intensive, as they usually require many iterations in the reverse diffusion process (RDP), making them impractical for streaming SE systems. In this paper, we propose to use discriminative scores from discriminative models in the first steps of the RDP. These discriminative scores require only one forward pass with the discriminative model for multiple RDP steps, thus greatly reducing computations. This approach also allows for performance improvements. We show that we can trade off between generative and discriminative capabilities as the number of steps with the discriminative score increases. Furthermore, we propose a novel streamable time-domain generative model with an algorithmic latency of 50 ms, which has no significant performance degradation compared to offline models.

eess.AS↗

arXiv2024

Finding Task-specific Subnetworks in Multi-task Spoken Language Understanding Model

Recently, multi-task spoken language understanding (SLU) models have emerged, designed to address various speech processing tasks. However, these models often rely on a large number of parameters. Also, they often encounter difficulties in adapting to new data for a specific task without experiencing catastrophic forgetting of previously trained tasks. In this study, we propose finding task-specific subnetworks within a multi-task SLU model via neural network pruning. In addition to model compression, we expect that the forgetting of previously trained tasks can be mitigated by updating only a task-specific subnetwork. We conduct experiments on top of the state-of-the-art multi-task SLU model ``UniverSLU'', trained for several tasks such as emotion recognition (ER), intent classification (IC), and automatic speech recognition (ASR). We show that pruned models were successful in adapting to additional ASR or IC data with minimal performance degradation on previously trained tasks.

cs.CL↗

arXiv2024

Rapid Language Adaptation for Multilingual E2E Speech Recognition Using Encoder Prompting

End-to-end multilingual speech recognition models handle multiple languages through a single model, often incorporating language identification to automatically detect the language of incoming speech. Since the common scenario is where the language is already known, these models can perform as language-specific by using language information as prompts, which is particularly beneficial for attention-based encoder-decoder architectures. However, the Connectionist Temporal Classification (CTC) approach, which enhances recognition via joint decoding and multi-task training, does not normally incorporate language prompts due to its conditionally independent output tokens. To overcome this, we introduce an encoder prompting technique within the self-conditioned CTC framework, enabling language-specific adaptation of the CTC model in a zero-shot manner. Our method has shown to significantly reduce errors by 28% on average and by 41% on low-resource languages.

cs.SD↗

arXiv2024

MMM: Multi-Layer Multi-Residual Multi-Stream Discrete Speech Representation from Self-supervised Learning Model

Speech discrete representation has proven effective in various downstream applications due to its superior compression rate of the waveform, fast convergence during training, and compatibility with other modalities. Discrete units extracted from self-supervised learning (SSL) models have emerged as a prominent approach for obtaining speech discrete representation. However, while discrete units have shown effectiveness compared to spectral features, they still lag behind continuous SSL representations. In this work, we propose MMM, a multi-layer multi-residual multi-stream discrete units extraction method from SSL. Specifically, we introduce iterative residual vector quantization with K-means for different layers in an SSL model to extract multi-stream speech discrete representation. Through extensive experiments in speech recognition, speech resynthesis, and text-to-speech, we demonstrate the proposed MMM can surpass or on-par with neural codec's performance under various conditions.

cs.SD↗

arXiv2024

On the Evaluation of Speech Foundation Models for Spoken Language Understanding

The Spoken Language Understanding Evaluation (SLUE) suite of benchmark tasks was recently introduced to address the need for open resources and benchmarking of complex spoken language understanding (SLU) tasks, including both classification and sequence generation tasks, on natural speech. The benchmark has demonstrated preliminary success in using pre-trained speech foundation models (SFM) for these SLU tasks. However, the community still lacks a fine-grained understanding of the comparative utility of different SFMs. Inspired by this, we ask: which SFMs offer the most benefits for these complex SLU tasks, and what is the most effective approach for incorporating these SFMs? To answer this, we perform an extensive evaluation of multiple supervised and self-supervised SFMs using several evaluation protocols: (i) frozen SFMs with a lightweight prediction head, (ii) frozen SFMs with a complex prediction head, and (iii) fine-tuned SFMs with a lightweight prediction head. Although the supervised SFMs are pre-trained on much more speech recognition data (with labels), they do not always outperform self-supervised SFMs; the latter tend to perform at least as well as, and sometimes better than, supervised SFMs, especially on the sequence generation tasks in SLUE. While there is no universally optimal way of incorporating SFMs, the complex prediction head gives the best performance for most tasks, although it increases the inference time. We also introduce an open-source toolkit and performance leaderboard, SLUE-PERB, for these tasks and modeling strategies.

cs.CL↗

arXiv2024

ESPnet-SPK: full pipeline speaker embedding toolkit with reproducible recipes, self-supervised front-ends, and off-the-shelf models

This paper introduces ESPnet-SPK, a toolkit designed with several objectives for training speaker embedding extractors. First, we provide an open-source platform for researchers in the speaker recognition community to effortlessly build models. We provide several models, ranging from x-vector to recent SKA-TDNN. Through the modularized architecture design, variants can be developed easily. We also aspire to bridge developed models with other domains, facilitating the broad research community to effortlessly incorporate state-of-the-art embedding extractors. Pre-trained embedding extractors can be accessed in an off-the-shelf manner and we demonstrate the toolkit's versatility by showcasing its integration with two tasks. Another goal is to integrate with diverse self-supervised learning features. We release a reproducible recipe that achieves an equal error rate of 0.39% on the Vox1-O evaluation protocol using WavLM-Large with ECAPA-TDNN.

cs.SD↗

arXiv2024

Can you Remove the Downstream Model for Speaker Recognition with Self-Supervised Speech Features?

Self-supervised features are typically used in place of filter-bank features in speaker verification models. However, these models were originally designed to ingest filter-bank features as inputs, and thus, training them on top of self-supervised features assumes that both feature types require the same amount of learning for the task. In this work, we observe that pre-trained self-supervised speech features inherently include information required for downstream speaker verification task, and therefore, we can simplify the downstream model without sacrificing performance. To this end, we revisit the design of the downstream model for speaker verification using self-supervised features. We show that we can simplify the model to use 97.51% fewer parameters while achieving a 29.93% average improvement in performance on SUPERB. Consequently, we show that the simplified downstream model is more data efficient compared to baseline--it achieves better performance with only 60% of the training data.

cs.SD↗

arXiv2024

On the Effects of Heterogeneous Data Sources on Speech-to-Text Foundation Models

The Open Whisper-style Speech Model (OWSM) series was introduced to achieve full transparency in building advanced speech-to-text (S2T) foundation models. To this end, OWSM models are trained on 25 public speech datasets, which are heterogeneous in multiple ways. In this study, we advance the OWSM series by introducing OWSM v3.2, which improves on prior models by investigating and addressing the impacts of this data heterogeneity. Our study begins with a detailed analysis of each dataset, from which we derive two key strategies: data filtering with proxy task to enhance data quality, and the incorporation of punctuation and true-casing using an open large language model (LLM). With all other configurations staying the same, OWSM v3.2 improves performance over the OWSM v3.1 baseline while using 15% less training data.

cs.CL↗

arXiv2024

DiscreteSLU: A Large Language Model with Self-Supervised Discrete Speech Units for Spoken Language Understanding

The integration of pre-trained text-based large language models (LLM) with speech input has enabled instruction-following capabilities for diverse speech tasks. This integration requires the use of a speech encoder, a speech adapter, and an LLM, trained on diverse tasks. We propose the use of discrete speech units (DSU), rather than continuous-valued speech encoder outputs, that are converted to the LLM token embedding space using the speech adapter. We generate DSU using a self-supervised speech encoder followed by k-means clustering. The proposed model shows robust performance on speech inputs from seen/unseen domains and instruction-following capability in spoken question answering. We also explore various types of DSU extracted from different layers of the self-supervised speech encoder, as well as Mel frequency Cepstral Coefficients (MFCC). Our findings suggest that the ASR task and datasets are not crucial in instruction-tuning for spoken question answering tasks.

cs.CL↗

arXiv2024

Singing Voice Data Scaling-up: An Introduction to ACE-Opencpop and ACE-KiSing

In singing voice synthesis (SVS), generating singing voices from musical scores faces challenges due to limited data availability. This study proposes a unique strategy to address the data scarcity in SVS. We employ an existing singing voice synthesizer for data augmentation, complemented by detailed manual tuning, an approach not previously explored in data curation, to reduce instances of unnatural voice synthesis. This innovative method has led to the creation of two expansive singing voice datasets, ACE-Opencpop and ACE-KiSing, which are instrumental for large-scale, multi-singer voice synthesis. Through thorough experimentation, we establish that these datasets not only serve as new benchmarks for SVS but also enhance SVS performance on other singing voice datasets when used as supplementary resources. The corpora, pre-trained models, and their related training recipes are publicly available at ESPnet-Muskits (\url{https://github.com/espnet/espnet})

cs.SD↗

arXiv2024

Neural Blind Source Separation and Diarization for Distant Speech Recognition

This paper presents a neural method for distant speech recognition (DSR) that jointly separates and diarizes speech mixtures without supervision by isolated signals. A standard separation method for multi-talker DSR is a statistical multichannel method called guided source separation (GSS). While GSS does not require signal-level supervision, it relies on speaker diarization results to handle unknown numbers of active speakers. To overcome this limitation, we introduce and train a neural inference model in a weakly-supervised manner, employing the objective function of a statistical separation method. This training requires only multichannel mixtures and their temporal annotations of speaker activities. In contrast to GSS, the trained model can jointly separate and diarize speech mixtures without any auxiliary information. The experiments with the AMI corpus show that our method outperforms GSS with oracle diarization results regarding word error rates. The code is available online.

eess.AS↗