Towards Comprehensive Semantic Speech Embeddings for Chinese Dialects

Despite having hundreds of millions of speakers, Chinese dialects lag behind Mandarin in speech and language technologies. Most varieties are primarily spoken, making dialect-to-Mandarin speech-LLMs (large language models) more practical than dialect LLMs. Building dialect-to-Mandarin speech-LLMs requires speech representations with cross-dialect semantic alignment between Chinese dialects and Mandarin. In this paper, we achieve such a cross-dialect semantic alignment by training a speech encoder with ASR (automatic speech recognition)-only data, as demonstrated by speech-to-speech retrieval on a new benchmark of spoken Chinese varieties that we contribute. Our speech encoder further demonstrates state-of-the-art ASR performance on Chinese dialects. Together, our Chinese dialect benchmark, semantically aligned speech representations, and speech-to-speech retrieval evaluation lay the groundwork for future Chinese dialect speech-LLMs. We release the benchmark at https://github.com/kalvinchang/yubao.

Latent-Level Enhancement with Flow Matching for Robust Automatic Speech Recognition

Noise-robust automatic speech recognition (ASR) has been commonly addressed by applying speech enhancement (SE) at the waveform level before recognition. However, speech-level enhancement does not always translate into consistent recognition improvements due to residual distortions and mismatches with the latent space of the ASR encoder. In this letter, we introduce a complementary strategy termed latent-level enhancement, where distorted representations are refined during ASR inference. Specifically, we propose a plug-and-play Flow Matching Refinement module (FM-Refiner) that operates on the output latents of a pretrained CTC-based ASR encoder. Trained to map imperfect latents-either directly from noisy inputs or from enhanced-but-imperfect speech-toward their clean counterparts, the FM-Refiner is applied only at inference, without fine-tuning ASR parameters. Experiments show that FM-Refiner consistently reduces word error rate, both when directly applied to noisy inputs and when combined with conventional SE front-ends. These results demonstrate that latent-level refinement via flow matching provides a lightweight and effective complement to existing SE approaches for robust ASR.

WESR: Scaling and Evaluating Word-level Event-Speech Recognition

Speech conveys not only linguistic information but also rich non-verbal vocal events such as laughing and crying. While semantic transcription is well-studied, the precise localization of non-verbal events remains a critical yet under-explored challenge. Current methods suffer from insufficient task definitions with limited category coverage and ambiguous temporal granularity. They also lack standardized evaluation frameworks, hindering the development of downstream applications. To bridge this gap, we first develop a refined taxonomy of 21 vocal events, with a new categorization into discrete (standalone) versus continuous (mixed with speech) types. Based on the refined taxonomy, we introduce WESR-Bench, an expert-annotated evaluation set (900+ utterances) with a novel position-aware protocol that disentangles ASR errors from event detection, enabling precise localization measurement for both discrete and continuous events. We also build a strong baseline by constructing a 1,700+ hour corpus, and train specialized models, surpassing both open-source audio-language models and commercial APIs while preserving ASR quality. We anticipate that WESR will serve as a foundational resource for future research in modeling rich, real-world auditory scenes.

LLMs-Integrated Automatic Hate Speech Recognition Using Controllable Text Generation Models

This paper proposes an automatic speech recognition (ASR) model for hate speech using large language models (LLMs). The proposed method integrates the encoder of the ASR model with the decoder of the LLMs, enabling simultaneous transcription and censorship tasks to prevent the exposure of harmful content. Instruction tuning of the LLM to mask hate-related words with specific tokens requires an annotated hate speech dataset, which is limited. We generate text samples using an LLM with the Chain-of-Thought (CoT) prompting technique guided by cultural context and examples and then convert them into speech samples using a text-to-speech (TTS) system. However, some of them contain non-hate speech samples with hate-related words, which degrades the censorship performance. This paper filters the samples which text classification models correctly label as hate content. By adjusting the threshold for the number of correct answer models, we can control the level of hate in the generated dataset, allowing us to train the LLMs through curriculum learning in a gradual manner. Experimental results show that the proposed method achieves a masking accuracy of 58.6\% for hate-related words, surpassing previous baselines. We also confirm that the curriculum training contributes to the efficiency of both transcription and censorship tasks.

Stuttering-Aware Automatic Speech Recognition for Indonesian Language

Automatic speech recognition systems have achieved remarkable performance on fluent speech but continue to degrade significantly when processing stuttered speech, a limitation that is particularly acute for low-resource languages like Indonesian where specialized datasets are virtually non-existent. To overcome this scarcity, we propose a data augmentation framework that generates synthetic stuttered audio by injecting repetitions and prolongations into fluent text through a combination of rule-based transformations and large language models followed by text-to-speech synthesis. We apply this synthetic data to fine-tune a pre-trained Indonesian Whisper model using transfer learning, enabling the architecture to adapt to dysfluent acoustic patterns without requiring large-scale real-world recordings. Our experiments demonstrate that this targeted synthetic exposure consistently reduces recognition errors on stuttered speech while maintaining performance on fluent segments, validating the utility of synthetic data pipelines for developing more inclusive speech technologies in under-represented languages.

Task Arithmetic with Support Languages for Low-Resource ASR

The development of resource-constrained approaches to automatic speech recognition (ASR) is of great interest due to its broad applicability to many low-resource languages for which there is scant usable data. Existing approaches to many low-resource natural language processing tasks leverage additional data from higher-resource languages that are closely related to a target low-resource language. One increasingly popular approach uses task arithmetic to combine models trained on different tasks to create a model for a task where there is little to no training data. In this paper, we consider training on a particular language to be a task, and we generate task vectors by fine-tuning variants of the Whisper ASR system. For pairings of high- and low-resource languages, we merge task vectors via a linear combination, optimizing the weights of the linear combination on the downstream word error rate on the low-resource target language's validation set. We find that this approach consistently improves performance on the target languages.

Variational decomposition autoencoding improves disentanglement of latent representations

Understanding the structure of complex, nonstationary, high-dimensional time-evolving signals is a central challenge in scientific data analysis. In many domains, such as speech and biomedical signal processing, the ability to learn disentangled and interpretable representations is critical for uncovering latent generative mechanisms. Traditional approaches to unsupervised representation learning, including variational autoencoders (VAEs), often struggle to capture the temporal and spectral diversity inherent in such data. Here we introduce variational decomposition autoencoding (VDA), a framework that extends VAEs by incorporating a strong structural bias toward signal decomposition. VDA is instantiated through variational decomposition autoencoders (DecVAEs), i.e., encoder-only neural networks that combine a signal decomposition model, a contrastive self-supervised task, and variational prior approximation to learn multiple latent subspaces aligned with time-frequency characteristics. We demonstrate the effectiveness of DecVAEs on simulated data and three publicly available scientific datasets, spanning speech recognition, dysarthria severity evaluation, and emotional speech classification. Our results demonstrate that DecVAEs surpass state-of-the-art VAE-based methods in terms of disentanglement quality, generalization across tasks, and the interpretability of latent encodings. These findings suggest that decomposition-aware architectures can serve as robust tools for extracting structured representations from dynamic signals, with potential applications in clinical diagnostics, human-computer interaction, and adaptive neurotechnologies.

Multi-channel multi-speaker transformer for speech recognition

With the development of teleconferencing and in-vehicle voice assistants, far-field multi-speaker speech recognition has become a hot research topic. Recently, a multi-channel transformer (MCT) has been proposed, which demonstrates the ability of the transformer to model far-field acoustic environments. However, MCT cannot encode high-dimensional acoustic features for each speaker from mixed input audio because of the interference between speakers. Based on these, we propose the multi-channel multi-speaker transformer (M2Former) for far-field multi-speaker ASR in this paper. Experiments on the SMS-WSJ benchmark show that the M2Former outperforms the neural beamformer, MCT, dual-path RNN with transform-average-concatenate and multi-channel deep clustering based end-to-end systems by 9.2%, 14.3%, 24.9%, and 52.2% respectively, in terms of relative word error rate reduction.

An Intelligent AI glasses System with Multi-Agent Architecture for Real-Time Voice Processing and Task Execution

This paper presents an AI glasses system that integrates real-time voice processing, artificial intelligence(AI) agents, and cross-network streaming capabilities. The system employs dual-agent architecture where Agent 01 handles Automatic Speech Recognition (ASR) and Agent 02 manages AI processing through local Large Language Models (LLMs), Model Context Protocol (MCP) tools, and Retrieval-Augmented Generation (RAG). The system supports real-time RTSP streaming for voice and video data transmission, eye tracking data collection, and remote task execution through RabbitMQ messaging. Implementation demonstrates successful voice command processing with multilingual support and cross-platform task execution capabilities.

QMAVIS: Long Video-Audio Understanding using Fusion of Large Multimodal Models

Large Multimodal Models (LMMs) for video-audio understanding have traditionally been evaluated only on shorter videos of a few minutes long. In this paper, we introduce QMAVIS (Q Team-Multimodal Audio Video Intelligent Sensemaking), a novel long video-audio understanding pipeline built through a late fusion of LMMs, Large Language Models, and speech recognition models. QMAVIS addresses the gap in long-form video analytics, particularly for longer videos of a few minutes to beyond an hour long, opening up new potential applications in sensemaking, video content analysis, embodied AI, etc. Quantitative experiments using QMAVIS demonstrated a 38.75% improvement over state-of-the-art video-audio LMMs like VideoLlaMA2 and InternVL2 on the VideoMME (with subtitles) dataset, which comprises long videos with audio information. Evaluations on other challenging video understanding datasets like PerceptionTest and EgoSchema saw up to 2% improvement, indicating competitive performance. Qualitative experiments also showed that QMAVIS is able to extract the nuances of different scenes in a long video audio content while understanding the overarching narrative. Ablation studies were also conducted to ascertain the impact of each component in the fusion pipeline.