Wearable devices like smart glasses are approaching the compute capability to seamlessly generate real-time closed captions for live conversations. We build on our recently introduced directional Automatic Speech Recognition (ASR) for smart glasses that have microphone arrays, which fuses multi-channel ASR with serialized output training, for wearer/conversation-partner disambiguation as well as suppression of cross-talk speech from non-target directions and noise. When ASR work is part of a broader system-development process, one may be faced with changes to microphone geometries as system development progresses. This paper aims to make multi-channel ASR insensitive to limited variations of microphone-array geometry. We show that a model trained on multiple similar geometries is largely agnostic and generalizes well to new geometries, as long as they are not too different. Furthermore, training the model this way improves accuracy for seen geometries by 15 to 28\% relative. Lastly, we refine the beamforming by a novel Non-Linearly Constrained Minimum Variance criterion.
The primary goal of the L3DAS23 Signal Processing Grand Challenge at ICASSP 2023 is to promote and support collaborative research on machine learning for 3D audio signal processing, with a specific emphasis on 3D speech enhancement and 3D Sound Event Localization and Detection in Extended Reality applications. As part of our latest competition, we provide a brand-new dataset, which maintains the same general characteristics of the L3DAS21 and L3DAS22 datasets, but with first-order Ambisonics recordings from multiple reverberant simulated environments. Moreover, we start exploring an audio-visual scenario by providing images of these environments, as perceived by the different microphone positions and orientations. We also propose updated baseline models for both tasks that can now support audio-image couples as input and a supporting API to replicate our results. Finally, we present the results of the participants. Further details about the challenge are available at https://www.l3das.com/icassp2023.
Aphasia, a language disorder primarily caused by a stroke, is traditionally diagnosed using behavioral language tests. However, these tests are time-consuming, require manual interpretation by trained clinicians, suffer from low ecological validity, and diagnosis can be biased by comorbid motor and cognitive problems present in aphasia. In this study, we introduce an automated screening tool for speech processing impairments in aphasia that relies on time-locked brain responses to speech, known as neural tracking, within a deep learning framework. We modeled electroencephalography (EEG) responses to acoustic, segmentation, and linguistic speech representations of a story using convolutional neural networks trained on a large sample of healthy participants, serving as a model for intact neural tracking of speech. Subsequently, we evaluated our models on an independent sample comprising 26 individuals with aphasia (IWA) and 22 healthy controls. Our results reveal decreased tracking of all speech representations in IWA. Utilizing a support vector machine classifier with neural tracking measures as input, we demonstrate high accuracy in aphasia detection at the individual level (85.42\%) in a time-efficient manner (requiring 9 minutes of EEG data). Given its high robustness, time efficiency, and generalizability to unseen data, our approach holds significant promise for clinical applications.
Recently, instruction-following audio-language models have received broad attention for human-audio interaction. However, the absence of benchmarks capable of evaluating audio-centric interaction capabilities has impeded advancements in this field. Previous models primarily focus on assessing different fundamental tasks, such as Automatic Speech Recognition (ASR), and lack an assessment of the open-ended generative capabilities centered around audio. Thus, it is challenging to track the progression in the Large Audio-Language Models (LALMs) domain and to provide guidance for future improvement. In this paper, we introduce AIR-Bench (\textbf{A}udio \textbf{I}nst\textbf{R}uction \textbf{Bench}mark), the first benchmark designed to evaluate the ability of LALMs to understand various types of audio signals (including human speech, natural sounds, and music), and furthermore, to interact with humans in the textual format. AIR-Bench encompasses two dimensions: \textit{foundation} and \textit{chat} benchmarks. The former consists of 19 tasks with approximately 19k single-choice questions, intending to inspect the basic single-task ability of LALMs. The latter one contains 2k instances of open-ended question-and-answer data, directly assessing the comprehension of the model on complex audio and its capacity to follow instructions. Both benchmarks require the model to generate hypotheses directly. We design a unified framework that leverages advanced language models, such as GPT-4, to evaluate the scores of generated hypotheses given the meta-information of the audio. Experimental results demonstrate a high level of consistency between GPT-4-based evaluation and human evaluation. By revealing the limitations of existing LALMs through evaluation results, AIR-Bench can provide insights into the direction of future research.
Two-stage pipeline is popular in speech enhancement tasks due to its superiority over traditional single-stage methods. The current two-stage approaches usually enhance the magnitude spectrum in the first stage, and further modify the complex spectrum to suppress the residual noise and recover the speech phase in the second stage. The above whole process is performed in the short-time Fourier transform (STFT) spectrum domain. In this paper, we re-implement the above second sub-process in the short-time discrete cosine transform (STDCT) spectrum domain. The reason is that we have found STDCT performs greater noise suppression capability than STFT. Additionally, the implicit phase of STDCT ensures simpler and more efficient phase recovery, which is challenging and computationally expensive in the STFT-based methods. Therefore, we propose a novel two-stage framework called the STFT-STDCT spectrum fusion network (FDFNet) for speech enhancement in cross-spectrum domain. Experimental results demonstrate that the proposed FDFNet outperforms the previous two-stage methods and also exhibits superior performance compared to other advanced systems.
Existing studies on self-supervised speech representation learning have focused on developing new training methods and applying pre-trained models for different applications. However, the quality of these models is often measured by the performance of different downstream tasks. How well the representations access the information of interest is less studied. In this work, we take a closer look into existing self-supervised methods of speech from an information-theoretic perspective. We aim to develop metrics using mutual information to help practical problems such as model design and selection. We use linear probes to estimate the mutual information between the target information and learned representations, showing another insight into the accessibility to the target information from speech representations. Further, we explore the potential of evaluating representations in a self-supervised fashion, where we estimate the mutual information between different parts of the data without using any labels. Finally, we show that both supervised and unsupervised measures echo the performance of the models on layer-wise linear probing and speech recognition.
Cross-corpus speech emotion recognition (SER) aims to transfer emotional knowledge from a labeled source corpus to an unlabeled corpus. However, prior methods require access to source data during adaptation, which is unattainable in real-life scenarios due to data privacy protection concerns. This paper tackles a more practical task, namely source-free cross-corpus SER, where a pre-trained source model is adapted to the target domain without access to source data. To address the problem, we propose a novel method called emotion-aware contrastive adaptation network (ECAN). The core idea is to capture local neighborhood information between samples while considering the global class-level adaptation. Specifically, we propose a nearest neighbor contrastive learning to promote local emotion consistency among features of highly similar samples. Furthermore, relying solely on nearest neighborhoods may lead to ambiguous boundaries between clusters. Thus, we incorporate supervised contrastive learning to encourage greater separation between clusters representing different emotions, thereby facilitating improved class-level adaptation. Extensive experiments indicate that our proposed ECAN significantly outperforms state-of-the-art methods under the source-free cross-corpus SER setting on several speech emotion corpora.
Recent language model (LM) advancements have showcased impressive zero-shot voice conversion (VC) performance. However, existing LM-based VC models usually apply offline conversion from source semantics to acoustic features, demanding the complete source speech, and limiting their deployment to real-time applications. In this paper, we introduce StreamVoice, a novel streaming LM-based model for zero-shot VC, facilitating real-time conversion given arbitrary speaker prompts and source speech. Specifically, to enable streaming capability, StreamVoice employs a fully causal context-aware LM with a temporal-independent acoustic predictor, while alternately processing semantic and acoustic features at each time step of autoregression which eliminates the dependence on complete source speech. To address the potential performance degradation from the incomplete context in streaming processing, we enhance the context-awareness of the LM through two strategies: 1) teacher-guided context foresight, using a teacher model to summarize the present and future semantic context during training to guide the model's forecasting for missing context; 2) semantic masking strategy, promoting acoustic prediction from preceding corrupted semantic and acoustic input, enhancing context-learning ability. Notably, StreamVoice is the first LM-based streaming zero-shot VC model without any future look-ahead. Experimental results demonstrate StreamVoice's streaming conversion capability while maintaining zero-shot performance comparable to non-streaming VC systems.
Unsupervised learning objectives like language modeling and de-noising constitute a significant part in producing pre-trained models that perform various downstream applications from natural language understanding to conversational tasks. However, despite impressive conversational capabilities of recent large language model, their abilities to capture syntactic or semantic structure within text lag behind. We hypothesize that the mismatch between linguistic performance and competence in machines is attributable to insufficient transfer of linguistic structure knowledge to computational systems with currently popular pre-training objectives. We show that punctuation restoration transfers to improvements in in- and out-of-distribution performance on structure-related tasks like named entity recognition, open information extraction, chunking, and part-of-speech tagging. Punctuation restoration is an effective learning objective that can improve structure understanding and yield a more robust structure-aware representations of natural language.
In the evolving landscape of online communication, moderating hate speech (HS) presents an intricate challenge, compounded by the multimodal nature of digital content. This comprehensive survey delves into the recent strides in HS moderation, spotlighting the burgeoning role of large language models (LLMs) and large multimodal models (LMMs). Our exploration begins with a thorough analysis of current literature, revealing the nuanced interplay between textual, visual, and auditory elements in propagating HS. We uncover a notable trend towards integrating these modalities, primarily due to the complexity and subtlety with which HS is disseminated. A significant emphasis is placed on the advances facilitated by LLMs and LMMs, which have begun to redefine the boundaries of detection and moderation capabilities. We identify existing gaps in research, particularly in the context of underrepresented languages and cultures, and the need for solutions to handle low-resource settings. The survey concludes with a forward-looking perspective, outlining potential avenues for future research, including the exploration of novel AI methodologies, the ethical governance of AI in moderation, and the development of more nuanced, context-aware systems. This comprehensive overview aims to catalyze further research and foster a collaborative effort towards more sophisticated, responsible, and human-centric approaches to HS moderation in the digital era.\footnote{ \textcolor{red}{WARNING: This paper contains offensive examples.