Mahi Luthra, Jiayi Shen, Maxime Poli, Angelo Ortiz Tandazo, Yosuke Higuchi, Youssef Benchekroun, Martin Gleize, Charles-Éric Saint-James, Dongyan Lin, Phillip Rust, Angel Villar-Corrales, Surya, Vanessa Stark, Rashel Moritz, Juan Pino, Yann LeCun, Emmanuel Dupoux. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

We propose to utilize an instruction-tuned large language model (LLM) for guiding the text generation process in automatic speech recognition (ASR). Modern LLMs are adept at performing various text generation tasks through zero-shot learning, prompted with instructions designed for specific objectives. This paper explores the potential of LLMs to derive linguistic information that can facilitate text generation in end-to-end ASR models. Specifically, we instruct an LLM to correct grammatical errors in an ASR hypothesis and use the LLM-derived representations to refine the output further. The proposed model is built on the joint CTC and attention architecture, with the LLM serving as a front-end feature extractor for the decoder. The ASR hypothesis, subject to correction, is obtained from the encoder via CTC decoding and fed into the LLM along with a specific instruction. The decoder subsequently takes as input the LLM output to perform token predictions, combining acoustic information from the encoder and the powerful linguistic information provided by the LLM. Experimental results show that the proposed LLM-guided model achieves a relative gain of approximately 13% in word error rates across major benchmarks.

We propose an approach to detect parodies in singing voices, analyzing attention weights derived from an encoder-decoder-based automatic speech recognition (ASR) model. Here, parodies involve modifying and singing existing lyrics written for songs. Sharing such modified singing voices on the internet carries the potential risk of copyright infringement, posing the need of an automatic parody detection system. Given that songs typically comprise fixed lyrics, the pair of speech and its corresponding transcription can be used to analyze singing voices. In this work, we feed singing voices into an encoder-decoder-based ASR system and perform the decoding process using the corresponding lyrics in a teacher-forcing manner. Here, when the ASR model encounters a singing voice that includes a parody segment, there is a potential for the attention weights between the singing voice and the correct lyrics become collapsed. By identifying such misalignments in the attention weights, we attempt to detect parodies in singing voices. Experimental comparisons using real karaoke singing voice data demonstrate that the developed system achieves highly accurate parody detection performance by effectively identifying misalignments.

Much of the recent progress in automatic speech recognition (ASR) lies in developing an acoustic encoder, such as enlarging its capacity and designing a refined architecture for speech processing. With these highly optimized encoders, the decoder has become less influential in its role as a language model (LM). In this work, we explore an effective approach for employing the LM structure in an ASR model. The proposed Mask-Conformer augments a Conformer-based model with a mask-predict decoder, which learns output context via the masked LM objective. The mask-predict decoder is applied to stacks of encoder layers, where the decoder output explicitly conditions the subsequent layers using cross-attention. We also propose a fill-mask decoding algorithm that refines a sequence using the decoder’s linguistic information. Experimental results show that Mask-Conformer outperforms strong baselines on some tasks. In addition, our analyses validate the effectiveness of the proposed model design.

Attention-based encoder-decoder models with autoregressive (AR) decoding have proven to be the dominant approach for automatic speech recognition (ASR) due to their superior accuracy. However, they often suffer from slow inference. This is primarily attributed to the incremental calculation of the decoder. This work proposes a partially AR framework, which employs segment-level vectorized beam search for improving the inference speed of an ASR model based on the hybrid connectionist temporal classification (CTC) attention-based architecture. It first generates an initial hypothesis using greedy CTC decoding, identifying low-confidence tokens based on their output probabilities. We then utilize the decoder to perform segment-level vectorized beam search on these tokens, re-predicting in parallel with minimal decoder calculations. Experimental results show that our method is 12 to 13 times faster in inference on the LibriSpeech corpus over AR decoding whilst preserving high accuracy.

Brian Yan, Siddharth Dalmia, Yosuke Higuchi, Graham Neubig, Florian Metze, Alan W Black, Shinji Watanabe. Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics. 2023.

This paper presents InterMPL, a semi-supervised learning method of end-to-end automatic speech recognition (ASR) that performs pseudo-labeling (PL) with intermediate supervision. Momentum PL (MPL) trains a connectionist temporal classification (CTC)-based model on unlabeled data by continuously generating pseudo-labels on the fly and improving their quality. In contrast to autoregressive formulations, such as the attention-based encoder-decoder and transducer, CTC is well suited for MPL, or PL-based semi-supervised ASR in general, owing to its simple/fast inference algorithm and robustness against generating collapsed labels. However, CTC generally yields inferior performance than the autoregressive models due to the conditional independence assumption, thereby limiting the performance of MPL. We propose to enhance MPL by introducing intermediate loss, inspired by the recent advances in CTC-based modeling. Specifically, we focus on self-conditional and hierarchical conditional CTC, that apply auxiliary CTC losses to intermediate layers such that the conditional independence assumption is explicitly relaxed. We also explore how pseudo-labels should be generated and used as supervision for intermediate losses. Experimental results in different semi-supervised settings demonstrate that the proposed approach outperforms MPL and improves an ASR model by up to a 12.1% absolute performance gain. In addition, our detailed analysis validates the importance of the intermediate loss.

We present BERT-CTC-Transducer (BECTRA), a novel end-to-end automatic speech recognition (E2E-ASR) model formulated by the transducer with a BERT-enhanced encoder. Integrating a large-scale pre-trained language model (LM) into E2E-ASR has been actively studied, aiming to utilize versatile linguistic knowledge for generating accurate text. One crucial factor that makes this integration challenging lies in the vocabulary mismatch; the vocabulary constructed for a pre-trained LM is generally too large for E2E-ASR training and is likely to have a mismatch against a target ASR domain. To overcome such an issue, we propose BECTRA, an extended version of our previous BERT-CTC, that realizes BERT-based E2E-ASR using a vocabulary of interest. BECTRA is a transducer-based model, which adopts BERT-CTC for its encoder and trains an ASR-specific decoder using a vocabulary suitable for a target task. With the combination of the transducer and BERT-CTC, we also propose a novel inference algorithm for taking advantage of both autoregressive and non-autoregressive decoding. Experimental results on several ASR tasks, varying in amounts of data, speaking styles, and languages, demonstrate that BECTRA outperforms BERT-CTC by effectively dealing with the vocabulary mismatch while exploiting BERT knowledge.

While Transformers have achieved promising results in end-to-end (E2E) automatic speech recognition (ASR), their autoregressive (AR) structure becomes a bottleneck for speeding up the decoding process. For real-world deployment, ASR systems are desired to be highly accurate while achieving fast inference. Non-autoregressive (NAR) models have become a popular alternative due to their fast inference speed, but they still fall behind AR systems in recognition accuracy. To fulfill the two demands, in this paper, we propose a NAR CTC/attention model utilizing both pre-trained acoustic and language models: wav2vec2.0 and BERT. To bridge the modality gap between speech and text representations obtained from the pre-trained models, we design a novel modality conversion mechanism, which is more suitable for logographic languages. During inference, we employ a CTC branch to generate a target length, which enables the BERT to predict tokens in parallel. We also design a cache-based CTC/attention joint decoding method to improve the recognition accuracy while keeping the decoding speed fast. Experimental results show that the proposed NAR model greatly outperforms our strong wav2vec2.0 CTC baseline (15.1% relative CER reduction on AISHELL-1). The proposed NAR model significantly surpasses previous NAR systems on the AISHELL-1 benchmark and shows a potential for English tasks.

In end-to-end automatic speech recognition (ASR), a model is expected to implicitly learn representations suitable for recognizing a word-level sequence. However, the huge abstraction gap between input acoustic signals and output linguistic tokens makes it challenging for a model to learn the representations. In this work, to promote the word-level representation learning in end-to-end ASR, we propose a hierarchical conditional model that is based on connectionist temporal classification (CTC). Our model is trained by auxiliary CTC losses applied to intermediate layers, where the vocabulary size of each target subword sequence is gradually increased as the layer becomes close to the word-level output. Here, we make each level of sequence prediction explicitly conditioned on the previous sequences predicted at lower levels. With the proposed approach, we expect the proposed model to learn the word-level representations effectively by exploiting a hierarchy of linguistic structures. Experimental results on LibriSpeech-{100h, 960h} and TEDLIUM2 demonstrate that the proposed model improves over a standard CTCbased model and other competitive models from prior work. We further analyze the results to confirm the effectiveness of the intended representation learning with our model.

Pseudo-labeling (PL), a semi-supervised learning (SSL) method where a seed model performs self-training using pseudo-labels generated from untranscribed speech, has been shown to enhance the performance of end-to-end automatic speech recognition (ASR). Our prior work proposed momentum pseudo-labeling (MPL), which performs PL-based SSL via an interaction between online and offline models, inspired by the mean teacher framework. MPL achieves remarkable results on various semi-supervised settings, showing robustness to variations in the amount of data and domain mismatch severity. However, there is further room for improving the seed model used to initialize the MPL training, as it is in general critical for a PL-based method to start training from high-quality pseudo-labels. To this end, we propose to enhance MPL by (1) introducing the Conformer architecture to boost the overall recognition accuracy and (2) exploiting iterative pseudo-labeling with a language model to improve the seed model before applying MPL. The experimental results demonstrate that the proposed approaches effectively improve MPL performance, outperforming other PL-based methods. We also present in-depth investigations to make our improvements effective, e.g., with regard to batch normalization typically used in Conformer and LM quality.

Collecting sufficient labeled data for spoken language understanding (SLU) is expensive and time-consuming. Recent studies achieved promising results by using pre-trained models in low-resource scenarios. Inspired by this, we aim to ask: which (if any) pre-training strategies can improve performance across SLU benchmarks? To answer this question, we employ four types of pre-trained models and their combinations for SLU. We leverage self-supervised speech and language models (LM) pre-trained on large quantities of un-paired data to extract strong speech and text representations. We also explore using supervised models pre-trained on larger external automatic speech recognition (ASR) or SLU corpora. We conduct extensive experiments on the SLU Evaluation (SLUE) benchmark and observe self-supervised pre-trained models to be more powerful, with pre-trained LM and speech models being most beneficial for the Sentiment Analysis and Named Entity Recognition task, respectively.11Our code and models will be publicly available as part of the ESPnet-SLU toolkit.

Non-autoregressive (NAR) models simultaneously generate multiple outputs in a sequence, which significantly reduces the inference speed at the cost of accuracy drop compared to autoregressive baselines. Showing great potential for real-time applications, an increasing number of NAR models have been explored in different fields to mitigate the performance gap against AR models. In this work, we conduct a comparative study of various NAR modeling methods for end-to-end automatic speech recognition (ASR). Experiments are performed in the state-of-the-art setting using ESPnet. The results on various tasks provide interesting findings for developing an understanding of NAR ASR, such as the accuracy-speed trade-off and robustness against long-form utterances. We also show that the techniques can be combined for further improvement and applied to NAR end-to-end speech translation. All the implementations are publicly available to encourage further research in NAR speech processing.

In this study, we present recent developments on ESPnet: End-to- End Speech Processing toolkit, which mainly involves a recently proposed architecture called Conformer, Convolution-augmented Transformer. This paper shows the results for a wide range of end- to-end speech processing applications, such as automatic speech recognition (ASR), speech translations (ST), speech separation (SS) and text-to-speech (TTS). Our experiments reveal various training tips and significant performance benefits obtained with the Conformer on different tasks. These results are competitive or even outperform the current state-of-art Transformer models. We are preparing to release all-in-one recipes using open source and publicly available corpora for all the above tasks with pre-trained models. Our aim for this work is to contribute to our research community by reducing the burden of preparing state-of-the-art research environments usually requiring high resources.

For real-world deployment of automatic speech recognition (ASR), the system is desired to be capable of fast inference while relieving the requirement of computational resources. The recently proposed end-to-end ASR system based on mask-predict with connectionist temporal classification (CTC), Mask-CTC, fulfills this demand by generating tokens in a non-autoregressive fashion. While Mask-CTC achieves remarkably fast inference speed, its recognition performance falls behind that of conventional autoregressive (AR) systems. To boost the performance of Mask-CTC, we first propose to enhance the encoder network architecture by employing a recently proposed architecture called Conformer. Next, we propose new training and decoding methods by introducing auxiliary objective to predict the length of a partial target sequence, which allows the model to delete or insert tokens during inference. Experimental results on different ASR tasks show that the proposed approaches improve Mask-CTC significantly, outperforming a standard CTC model (15.5% → 9.1% WER on WSJ). Moreover, Mask-CTC now achieves competitive results to AR models with no degradation of inference speed (< 0.1 RTF using CPU). We also show a potential application of Mask-CTC to end-to-end speech translation.

Fast inference speed is an important goal towards real-world deployment of speech translation (ST) systems. End-to-end (E2E) models based on the encoder-decoder architecture are more suitable for this goal than traditional cascaded systems, but their effectiveness regarding decoding speed has not been explored so far. Inspired by recent progress in non-autoregressive (NAR) methods in text-based translation, which generates target tokens in parallel by eliminating conditional dependencies, we study the problem of NAR decoding for E2E-ST. We propose a novel NAR E2E-ST framework, Orthros, in which both NAR and autoregressive (AR) decoders are jointly trained on the shared speech encoder. The latter is used for selecting better translation among various length candidates generated from the former, which dramatically improves the effectiveness of a large length beam with negligible overhead. We further investigate effective length prediction methods from speech inputs and the impact of vocabulary sizes. Experiments on four benchmarks show the effectiveness of the proposed method in improving inference speed while maintaining competitive translation quality compared to state-of-the-art AR E2E-ST systems.

Pseudo-labeling (PL) has been shown to be effective in semisupervised automatic speech recognition (ASR), where a base model is self-trained with pseudo-labels generated from unlabeled data.While PL can be further improved by iteratively updating pseudo-labels as the model evolves, most of the previous approaches involve inefficient retraining of the model or intricate control of the label update.We present momentum pseudo-labeling (MPL), a simple yet effective strategy for semisupervised ASR.MPL consists of a pair of online and offline models that interact and learn from each other, inspired by the mean teacher method.The online model is trained to predict pseudo-labels generated on the fly by the offline model.The offline model maintains a momentum-based moving average of the online model.MPL is performed in a single training process and the interaction between the two models effectively helps them reinforce each other to improve the ASR performance.We apply MPL to an end-to-end ASR model based on the connectionist temporal classification.The experimental results demonstrate that MPL effectively improves over the base model and is scalable to different semi-supervised scenarios with varying amounts of data or domain mismatch.

We present our work on training speaker embeddings, especially effective for speaker diarization. For various speaker recognition tasks, extracting speaker embeddings using Deep Neural Networks (DNNs) has become major methods. These embeddings are generally trained to be discriminate speakers and be robust with respect to different acoustic conditions. In speaker diarization, however, the acoustic conditions can be used as consistent information for discriminating speakers. Such information can include the distances to a microphone in a meeting, or the channels for each speaker in telephone conversation recorded in monaural. Hence, the proposed speaker-embedding network leverages differences in acoustic conditions to train effective speaker embeddings for speaker diarization. The information on acoustic conditions can be anything that contributes to distinguishing between recording environments; for example, we explore using i-vectors. Experiments conducted on a practical diarization system demonstrated that the proposed embeddings significantly improve performance over embeddings without information on acoustic conditions.

We present Mask CTC, a novel non-autoregressive end-to-end automatic speech recognition (ASR) framework, which generates a sequence by refining outputs of the connectionist temporal classification (CTC).Neural sequence-to-sequence models are usually autoregressive: each output token is generated by conditioning on previously generated tokens, at the cost of requiring as many iterations as the output length.On the other hand, non-autoregressive models can simultaneously generate tokens within a constant number of iterations, which results in significant inference time reduction and better suits end-toend ASR model for real-world scenarios.In this work, Mask CTC model is trained using a Transformer encoder-decoder with joint training of mask prediction and CTC.During inference, the target sequence is initialized with the greedy CTC outputs and low-confidence tokens are masked based on the CTC probabilities.Based on the conditional dependence between output tokens, these masked low-confidence tokens are then predicted conditioning on the high-confidence tokens.Experimental results on different speech recognition tasks show that Mask CTC outperforms the standard CTC model (e.g., 17.9% → 12.1% WER on WSJ) and approaches the autoregressive model, requiring much less inference time using CPUs (0.07 RTF in Python implementation).All of our codes are publicly available at https://github.com/espnet/espnet.