Dialogue systems based on large language models (LLMs) have advanced significantly in recent years. However, dialectal variation remains a major challenge, particularly for systems that process spoken input. LLM-based speech language models (SLMs), which integrate LLMs with speech processing components, show promise for spoken language tasks, yet their ability to comprehend dialects has not been sufficiently studied. Moreover, it remains unclear how the dialectal understanding of the base LLM affects SLM performance. This study investigates the dialectal robustness of both LLMs and SLMs using Japanese dialects as a test case. We define robustness as the ratio of performance on dialectal versus standard inputs, enabling fair comparisons. Our experiments show that SLM robustness correlates with that of their text-based counterparts. Furthermore, training with dialectal data and fine-tuning the speech encoder each improves robustness in SLMs.

This paper presents a unified multi-speaker encoder (UME), a novel architecture that jointly learns representations for speaker diarization (SD), speech separation (SS), and multi-speaker automatic speech recognition (ASR) tasks using a shared speech foundational encoder. We leverage the hidden representations from multiple layers of UME as a residual weighted-sum encoding (RWSE) to effectively use information from different semantic levels, contributing to bottom-up alignment between tasks. This joint training approach captures the inherent inter-dependencies among the tasks, enhancing overall performance on overlapping speech data. Our evaluations demonstrate that UME substantially improves over the single-task baselines dedicated to SD, SS, and multi-speaker ASR on LibriMix evaluation sets. Notably, for SD, UME outperforms the previous studies, achieving diarization error rates of 1.37% and 2.29% on Libri2Mix and Libri3Mix evaluation sets, respectively.

Prompts are crucial for task definition and for improving the performance of large language models (LLM)-based systems. However, existing LLM-based multi-talker (MT) automatic speech recognition (ASR) systems either omit prompts or rely on simple task-definition prompts, with no prior work exploring the design of prompts to enhance performance. In this paper, we propose extracting serialized output prompts (SOP) and explicitly guiding the LLM using structured prompts to improve system performance (SOP-MT-ASR). A Separator and serialized Connectionist Temporal Classification (CTC) layers are inserted after the speech encoder to separate and extract MT content from the mixed speech encoding in a first-speaking-first-out manner. Subsequently, the SOP, which serves as a prompt for LLMs, is obtained by decoding the serialized CTC outputs using greedy search. To train the model effectively, we design a threestage training strategy, consisting of serialized output training (SOT) fine-tuning, serialized speech information extraction, and SOP-based adaptation. Experimental results on the LibriMix dataset show that, although the LLM-based SOT model performs well in the two-talker scenario, it fails to fully leverage LLMs under more complex conditions, such as the three-talker scenario. The proposed SOP approach significantly improved performance under both two- and three-talker conditions.

There has been an increasing interest in large speech models that can perform multiple tasks in a single model.Such models usually adopt an encoder-decoder or decoder-only architecture due to their popularity and good performance in many domains.However, autoregressive models can be slower during inference compared to non-autoregressive models and also have potential risks of hallucination.Though prior studies observed promising results of non-autoregressive models for certain tasks at small scales, it remains unclear if they can be scaled to speech-to-text generation in diverse languages and tasks.Inspired by the Open Whisper-style Speech Model (OWSM) project, we propose OWSM-CTC, a novel encoder-only speech foundation model based on Connectionist Temporal Classification (CTC).It is trained on 180k hours of public audio data for multilingual automatic speech recognition (ASR), speech translation (ST), and language identification (LID).Compared to encoder-decoder OWSM, our OWSM-CTC achieves competitive results on ASR and up to 24% relative improvement on ST, while it is more robust and 3 to 4 times faster for inference.OWSM-CTC also improves the longform ASR result with 20x speed-up.We will publicly release our code, pre-trained model, and training logs to promote open science in speech foundation models. 1

End-to-end (E2E) automatic speech recognition (ASR) methods exhibit remarkable performance. However, since the performance of such methods is intrinsically linked to the context present in the training data, E2E-ASR methods do not perform as desired for unseen user contexts (e.g., technical terms, personal names, and playlists). Thus, E2E-ASR methods must be easily contextualized by the user or developer. This paper proposes an attention-based contextual biasing method that can be customized using an editable phrase list (referred to as a bias list). The proposed method can be trained effectively by combining a bias phrase index loss and special tokens to detect the bias phrases in the input speech data. In addition, to improve the contextualization performance during inference further, we propose a bias phrase boosted (BPB) beam search algorithm based on the bias phrase index probability. Experimental results demonstrate that the proposed method consistently improves the word error rate and the character error rate of the target phrases in the bias list on both the Librispeech-960 (English) and our in-house (Japanese) dataset, respectively.

Deep biasing (DB) enhances the performance of end-to-end automatic speech recognition (E2E-ASR) models for rare words or contextual phrases using a bias list. However, most existing methods treat bias phrases as sequences of subwords in a predefined static vocabulary. This naive sequence decomposition produces unnatural token patterns, significantly lowering their occurrence probability. More advanced techniques address this problem by expanding the vocabulary with additional modules, including the external language model shallow fusion or rescoring. However, they result in increasing the workload due to the additional modules. This paper proposes a dynamic vocabulary where bias tokens can be added during inference. Each entry in a bias list is represented as a single token, unlike a sequence of existing subword tokens. This approach eliminates the need to learn subword dependencies within the bias phrases. This method is easily applied to various architectures because it only expands the embedding and output layers in common E2E-ASR architectures. Experimental results demonstrate that the proposed method improves the bias phrase WER on English and Japanese datasets by $3.1-4.9$ points compared with the conventional DB method.

Pre-training speech models on large volumes of data has achieved remarkable success. OpenAI Whisper is a multilingual multitask model trained on 680k hours of supervised speech data. It generalizes well to various speech recognition and translation benchmarks even in a zero-shot setup. However, the full pipeline for developing such models (from data collection to training) is not publicly accessible, which makes it difficult for researchers to further improve its performance and address training-related issues such as efficiency, robustness, fairness, and bias. This work presents an Open Whisper-style Speech Model (OWSM), which reproduces Whisperstyle training using an open-source toolkit and publicly available data. OWSM even supports more translation directions and can be more efficient to train. We will publicly release all scripts used for data preparation, training, inference, and scoring as well as pretrained models and training logs to promote open science.11https://github.com/espnet/espnet

This paper proposes an online adaptation method for Fourier series based acoustic transfer function (TF) models for robot audition systems based on microphone array signal processing. The TF represents the signal propagation characteristics from a sound source to a microphone, which is an essential component for real-world auditory scene analysis, including sound source localization and separation. The real-world applications of TF-based array signal processing requires two characteristics: 1) adaptability to changes in the acoustic environment (changes in the signal propagation characteristics between the sound source and the microphone), and 2) a lightweight TF set for use in embedded systems such as robots with limited memory and computational resources. This paper proposes an online adaptation method for lightweight TF models using the Fourier series expansion. This method has both above two characteristics. Experimental results showed that the use of TF set adapted online using the proposed method performs better sound source localization and separation performance than existing online TF adaptation methods.

Abstract This paper proposes a multichannel environmental sound segmentation method. Environmental sound segmentation is an integrated method to achieve sound source localization, sound source separation and classification, simultaneously. When multiple microphones are available, spatial features can be used to improve the localization and separation accuracy of sounds from different directions; however, conventional methods have three drawbacks: (a) Sound source localization and sound source separation methods using spatial features and classification using spectral features trained in the same neural network, may overfit to the relationship between the direction of arrival and the class of a sound, thereby reducing their reliability to deal with novel events. (b) Although permutation invariant training used in autonomous speech recognition could be extended, it is impractical for environmental sounds that include an unlimited number of sound sources. (c) Various features, such as complex values of short time Fourier transform and interchannel phase differences have been used as spatial features, but no study has compared them. This paper proposes a multichannel environmental sound segmentation method comprising two discrete blocks, a sound source localization and separation block and a sound source separation and classification block. By separating the blocks, overfitting to the relationship between the direction of arrival and the class is avoided. Simulation experiments using created datasets including 75-class environmental sounds showed the root mean squared error of the proposed method was lower than that of conventional methods.

This paper proposes an environmental sound segmentation method using Mask U-Net. Recent research in robot audition has analyzed noise reduction, section detection, and sound source separation for use in a real-world environment with many noises and overlaps. However, conventional methods apply respective functions in cascades. The biggest problem of cascade systems is the accumulation of errors generated at each function block. Although many methods of human voice separation have been proposed, robots operating in a real-world environment must be able to separate not only human voices but other environmental sounds. Unlike traditional sound source separation using spatial information, environmental sound segmentation must simultaneously detect sections and separate sound sources based on pre-trained features. One such method, U-Net, which was proposed for semantic segmentation of images, has been applied to the separation of singing voices. However, this method deals only with limited classes of sounds. The current study proposes an environmental sound segmentation method using Mask U-Net, which combines segmentation using U-Net with sound event detection using CNN to 75-classes of environmental sounds. Experimental application confirmed that this method improved learning speed and sound source separation compared with the conventional method.