2024-05-14 更新
Dual-path Mamba: Short and Long-term Bidirectional Selective Structured State Space Models for Speech Separation
Authors:Xilin Jiang, Cong Han, Nima Mesgarani
Transformers have been the most successful architecture for various speech modeling tasks, including speech separation. However, the self-attention mechanism in transformers with quadratic complexity is inefficient in computation and memory. Recent models incorporate new layers and modules along with transformers for better performance but also introduce extra model complexity. In this work, we replace transformers with Mamba, a selective state space model, for speech separation. We propose dual-path Mamba, which models short-term and long-term forward and backward dependency of speech signals using selective state spaces. Our experimental results on the WSJ0-2mix data show that our dual-path Mamba models of comparably smaller sizes outperform state-of-the-art RNN model DPRNN, CNN model WaveSplit, and transformer model Sepformer. Code: https://github.com/xi-j/Mamba-TasNet
PDF work in progress
点此查看论文截图
Leveraging Speech for Gesture Detection in Multimodal Communication
Authors:Esam Ghaleb, Ilya Burenko, Marlou Rasenberg, Wim Pouw, Ivan Toni, Peter Uhrig, Anna Wilson, Judith Holler, Aslı Özyürek, Raquel Fernández
Gestures are inherent to human interaction and often complement speech in face-to-face communication, forming a multimodal communication system. An important task in gesture analysis is detecting a gesture’s beginning and end. Research on automatic gesture detection has primarily focused on visual and kinematic information to detect a limited set of isolated or silent gestures with low variability, neglecting the integration of speech and vision signals to detect gestures that co-occur with speech. This work addresses this gap by focusing on co-speech gesture detection, emphasising the synchrony between speech and co-speech hand gestures. We address three main challenges: the variability of gesture forms, the temporal misalignment between gesture and speech onsets, and differences in sampling rate between modalities. We investigate extended speech time windows and employ separate backbone models for each modality to address the temporal misalignment and sampling rate differences. We utilize Transformer encoders in cross-modal and early fusion techniques to effectively align and integrate speech and skeletal sequences. The study results show that combining visual and speech information significantly enhances gesture detection performance. Our findings indicate that expanding the speech buffer beyond visual time segments improves performance and that multimodal integration using cross-modal and early fusion techniques outperforms baseline methods using unimodal and late fusion methods. Additionally, we find a correlation between the models’ gesture prediction confidence and low-level speech frequency features potentially associated with gestures. Overall, the study provides a better understanding and detection methods for co-speech gestures, facilitating the analysis of multimodal communication.
PDF
点此查看论文截图
Improving Membership Inference in ASR Model Auditing with Perturbed Loss Features
Authors:Francisco Teixeira, Karla Pizzi, Raphael Olivier, Alberto Abad, Bhiksha Raj, Isabel Trancoso
Membership Inference (MI) poses a substantial privacy threat to the training data of Automatic Speech Recognition (ASR) systems, while also offering an opportunity to audit these models with regard to user data. This paper explores the effectiveness of loss-based features in combination with Gaussian and adversarial perturbations to perform MI in ASR models. To the best of our knowledge, this approach has not yet been investigated. We compare our proposed features with commonly used error-based features and find that the proposed features greatly enhance performance for sample-level MI. For speaker-level MI, these features improve results, though by a smaller margin, as error-based features already obtained a high performance for this task. Our findings emphasise the importance of considering different feature sets and levels of access to target models for effective MI in ASR systems, providing valuable insights for auditing such models.
PDF Trustworthy Speech Processing, Satellite Workshop at ICASSP 2024
点此查看论文截图
CoS: Enhancing Personalization and Mitigating Bias with Context Steering
Authors:Jerry Zhi-Yang He, Sashrika Pandey, Mariah L. Schrum, Anca Dragan
When querying a large language model (LLM), the context, i.e. personal, demographic, and cultural information specific to an end-user, can significantly shape the response of the LLM. For example, asking the model to explain Newton’s second law with the context “I am a toddler” yields a different answer compared to the context “I am a physics professor.” Proper usage of the context enables the LLM to generate personalized responses, whereas inappropriate contextual influence can lead to stereotypical and potentially harmful generations (e.g. associating “female” with “housekeeper”). In practice, striking the right balance when leveraging context is a nuanced and challenging problem that is often situation-dependent. One common approach to address this challenge is to fine-tune LLMs on contextually appropriate responses. However, this approach is expensive, time-consuming, and not controllable for end-users in different situations. In this work, we propose Context Steering (CoS) - a simple training-free method that can be easily applied to autoregressive LLMs at inference time. By measuring the contextual influence in terms of token prediction likelihood and modulating it, our method enables practitioners to determine the appropriate level of contextual influence based on their specific use case and end-user base. We showcase a variety of applications of CoS including amplifying the contextual influence to achieve better personalization and mitigating unwanted influence for reducing model bias. In addition, we show that we can combine CoS with Bayesian Inference to quantify the extent of hate speech on the internet. We demonstrate the effectiveness of CoS on state-of-the-art LLMs and benchmarks.
PDF
点此查看论文截图
Real-time multichannel deep speech enhancement in hearing aids: Comparing monaural and binaural processing in complex acoustic scenarios
Authors:Nils L. Westhausen, Hendrik Kayser, Theresa Jansen, Bernd T. Meyer
Deep learning has the potential to enhance speech signals and increase their intelligibility for users of hearing aids. Deep models suited for real-world application should feature a low computational complexity and low processing delay of only a few milliseconds. In this paper, we explore deep speech enhancement that matches these requirements and contrast monaural and binaural processing algorithms in two complex acoustic scenes. Both algorithms are evaluated with objective metrics and in experiments with hearing-impaired listeners performing a speech-in-noise test. Results are compared to two traditional enhancement strategies, i.e., adaptive differential microphone processing and binaural beamforming. While in diffuse noise, all algorithms perform similarly, the binaural deep learning approach performs best in the presence of spatial interferers. Through a post-analysis, this can be attributed to improvements at low SNRs and to precise spatial filtering.
PDF This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible
点此查看论文截图
TIPAA-SSL: Text Independent Phone-to-Audio Alignment based on Self-Supervised Learning and Knowledge Transfer
Authors:Noé Tits, Prernna Bhatnagar, Thierry Dutoit
In this paper, we present a novel approach for text independent phone-to-audio alignment based on phoneme recognition, representation learning and knowledge transfer. Our method leverages a self-supervised model (wav2vec2) fine-tuned for phoneme recognition using a Connectionist Temporal Classification (CTC) loss, a dimension reduction model and a frame-level phoneme classifier trained thanks to forced-alignment labels (using Montreal Forced Aligner) to produce multi-lingual phonetic representations, thus requiring minimal additional training. We evaluate our model using synthetic native data from the TIMIT dataset and the SCRIBE dataset for American and British English, respectively. Our proposed model outperforms the state-of-the-art (charsiu) in statistical metrics and has applications in language learning and speech processing systems. We leave experiments on other languages for future work but the design of the system makes it easily adaptable to other languages.
PDF
点此查看论文截图
MMGER: Multi-modal and Multi-granularity Generative Error Correction with LLM for Joint Accent and Speech Recognition
Authors:Bingshen Mu, Yangze Li, Qijie Shao, Kun Wei, Xucheng Wan, Naijun Zheng, Huan Zhou, Lei Xie
Despite notable advancements in automatic speech recognition (ASR), performance tends to degrade when faced with adverse conditions. Generative error correction (GER) leverages the exceptional text comprehension capabilities of large language models (LLM), delivering impressive performance in ASR error correction, where N-best hypotheses provide valuable information for transcription prediction. However, GER encounters challenges such as fixed N-best hypotheses, insufficient utilization of acoustic information, and limited specificity to multi-accent scenarios. In this paper, we explore the application of GER in multi-accent scenarios. Accents represent deviations from standard pronunciation norms, and the multi-task learning framework for simultaneous ASR and accent recognition (AR) has effectively addressed the multi-accent scenarios, making it a prominent solution. In this work, we propose a unified ASR-AR GER model, named MMGER, leveraging multi-modal correction, and multi-granularity correction. Multi-task ASR-AR learning is employed to provide dynamic 1-best hypotheses and accent embeddings. Multi-modal correction accomplishes fine-grained frame-level correction by force-aligning the acoustic features of speech with the corresponding character-level 1-best hypothesis sequence. Multi-granularity correction supplements the global linguistic information by incorporating regular 1-best hypotheses atop fine-grained multi-modal correction to achieve coarse-grained utterance-level correction. MMGER effectively mitigates the limitations of GER and tailors LLM-based ASR error correction for the multi-accent scenarios. Experiments conducted on the multi-accent Mandarin KeSpeech dataset demonstrate the efficacy of MMGER, achieving a 26.72% relative improvement in AR accuracy and a 27.55% relative reduction in ASR character error rate, compared to a well-established standard baseline.
PDF
点此查看论文截图
Speaker Characterization by means of Attention Pooling
Authors:Federico Costa, Miquel India, Javier Hernando
State-of-the-art Deep Learning systems for speaker verification are commonly based on speaker embedding extractors. These architectures are usually composed of a feature extractor front-end together with a pooling layer to encode variable-length utterances into fixed-length speaker vectors. The authors have recently proposed the use of a Double Multi-Head Self-Attention pooling for speaker recognition, placed between a CNN-based front-end and a set of fully connected layers. This has shown to be an excellent approach to efficiently select the most relevant features captured by the front-end from the speech signal. In this paper we show excellent experimental results by adapting this architecture to other different speaker characterization tasks, such as emotion recognition, sex classification and COVID-19 detection.
PDF IberSpeech 2022
点此查看论文截图
Fine-grained Speech Sentiment Analysis in Chinese Psychological Support Hotlines Based on Large-scale Pre-trained Model
Authors:Zhonglong Chen, Changwei Song, Yining Chen, Jianqiang Li, Guanghui Fu, Yongsheng Tong, Qing Zhao
Suicide and suicidal behaviors remain significant challenges for public policy and healthcare. In response, psychological support hotlines have been established worldwide to provide immediate help to individuals in mental crises. The effectiveness of these hotlines largely depends on accurately identifying callers’ emotional states, particularly underlying negative emotions indicative of increased suicide risk. However, the high demand for psychological interventions often results in a shortage of professional operators, highlighting the need for an effective speech emotion recognition model. This model would automatically detect and analyze callers’ emotions, facilitating integration into hotline services. Additionally, it would enable large-scale data analysis of psychological support hotline interactions to explore psychological phenomena and behaviors across populations. Our study utilizes data from the Beijing psychological support hotline, the largest suicide hotline in China. We analyzed speech data from 105 callers containing 20,630 segments and categorized them into 11 types of negative emotions. We developed a negative emotion recognition model and a fine-grained multi-label classification model using a large-scale pre-trained model. Our experiments indicate that the negative emotion recognition model achieves a maximum F1-score of 76.96%. However, it shows limited efficacy in the fine-grained multi-label classification task, with the best model achieving only a 41.74% weighted F1-score. We conducted an error analysis for this task, discussed potential future improvements, and considered the clinical application possibilities of our study. All the codes are public available.
PDF
点此查看论文截图
Open Implementation and Study of BEST-RQ for Speech Processing
Authors:Ryan Whetten, Titouan Parcollet, Marco Dinarelli, Yannick Estève
Self-Supervised Learning (SSL) has proven to be useful in various speech tasks. However, these methods are generally very demanding in terms of data, memory, and computational resources. BERT-based Speech pre-Training with Random-projection Quantizer (BEST-RQ), is an SSL method that has shown great performance on Automatic Speech Recognition (ASR) while being simpler than other SSL methods, such as wav2vec 2.0. Despite BEST-RQ’s great performance, details are lacking in the original paper, such as the amount of GPU/TPU hours used in pre-training, and there is no official easy-to-use open-source implementation. Furthermore, BEST-RQ has not been evaluated on other downstream tasks aside from ASR and speech translation. In this work, we describe a re-implementation of a Random-projection quantizer and perform a preliminary study with a comparison to wav2vec 2.0 on four downstream tasks. We discuss the details and differences of our implementation. We show that a random projection quantizer can achieve similar downstream performance as wav2vec 2.0 while decreasing training time by over a factor of two.
PDF Accepted in IEEE ICASSP 2024 workshop on Self-supervision in Audio, Speech and Beyond (SASB 2024)
点此查看论文截图
BERP: A Blind Estimator of Room Acoustic and Physical Parameters for Single-Channel Noisy Speech Signals
Authors:Lijun Wang, Yixian Lu, Ziyan Gao, Kai Li, Jianqiang Huang, Yuntao Kong, Shogo Okada
Room acoustic parameters (RAPs) and room physical parameters ( RPPs) are essential metrics for parameterizing the room acoustical characteristics (RAC) of a sound field around a listener’s local environment, offering comprehensive indications for various applications. The current RAPs and RPPs estimation methods either fall short of covering broad real-world acoustic environments in the context of real background noise or lack universal frameworks for blindly estimating RAPs and RPPs from noisy single-channel speech signals, particularly sound source distances, direction-of-arrival (DOA) of sound sources, and occupancy levels. On the other hand, in this paper, we propose a novel universal blind estimation framework called the blind estimator of room acoustical and physical parameters (BERP), by introducing a new stochastic room impulse response (RIR) model, namely, the sparse stochastic impulse response (SSIR) model, and endowing the BERP with a unified encoder and multiple separate predictors to estimate RPPs and SSIR parameters in parallel. This estimation framework enables the computationally efficient and universal estimation of room parameters by solely using noisy single-channel speech signals. Finally, all the RAPs can be simultaneously derived from the RIRs synthesized from SSIR model with the estimated parameters. To evaluate the effectiveness of the proposed BERP and SSIR models, we compile a task-specific dataset from several publicly available datasets. The results reveal that the BERP achieves state-of-the-art (SOTA) performance. Moreover, the evaluation results pertaining to the SSIR RIR model also demonstrated its efficacy. The code is available on GitHub.
PDF Submitted to IEEE/ACM Transaction on Audio Speech and Language Processing (TASLP)
点此查看论文截图
Muting Whisper: A Universal Acoustic Adversarial Attack on Speech Foundation Models
Authors:Vyas Raina, Rao Ma, Charles McGhee, Kate Knill, Mark Gales
Recent developments in large speech foundation models like Whisper have led to their widespread use in many automatic speech recognition (ASR) applications. These systems incorporate special tokens' in their vocabulary, such as $\texttt{<endoftext>}$, to guide their language generation process. However, we demonstrate that these tokens can be exploited by adversarial attacks to manipulate the model's behavior. We propose a simple yet effective method to learn a universal acoustic realization of Whisper's $\texttt{<endoftext>}$ token, which, when prepended to any speech signal, encourages the model to ignore the speech and only transcribe the special token, effectivelymuting’ the model. Our experiments demonstrate that the same, universal 0.64-second adversarial audio segment can successfully mute a target Whisper ASR model for over 97\% of speech samples. Moreover, we find that this universal adversarial audio segment often transfers to new datasets and tasks. Overall this work demonstrates the vulnerability of Whisper models to `muting’ adversarial attacks, where such attacks can pose both risks and potential benefits in real-world settings: for example the attack can be used to bypass speech moderation systems, or conversely the attack can also be used to protect private speech data.
PDF
点此查看论文截图
DP-DyLoRA: Fine-Tuning Transformer-Based Models On-Device under Differentially Private Federated Learning using Dynamic Low-Rank Adaptation
Authors:Jie Xu, Karthikeyan Saravanan, Rogier van Dalen, Haaris Mehmood, David Tuckey, Mete Ozay
Federated learning (FL) allows clients in an Internet of Things (IoT) system to collaboratively train a global model without sharing their local data with a server. However, clients’ contributions to the server can still leak sensitive information. Differential privacy (DP) addresses such leakage by providing formal privacy guarantees, with mechanisms that add randomness to the clients’ contributions. The randomness makes it infeasible to train large transformer-based models, common in modern IoT systems. In this work, we empirically evaluate the practicality of fine-tuning large scale on-device transformer-based models with differential privacy in a federated learning system. We conduct comprehensive experiments on various system properties for tasks spanning a multitude of domains: speech recognition, computer vision (CV) and natural language understanding (NLU). Our results show that full fine-tuning under differentially private federated learning (DP-FL) generally leads to huge performance degradation which can be alleviated by reducing the dimensionality of contributions through parameter-efficient fine-tuning (PEFT). Our benchmarks of existing DP-PEFT methods show that DP-Low-Rank Adaptation (DP-LoRA) consistently outperforms other methods. An even more promising approach, DyLoRA, which makes the low rank variable, when naively combined with FL would straightforwardly break differential privacy. We therefore propose an adaptation method that can be combined with differential privacy and call it DP-DyLoRA. Finally, we are able to reduce the accuracy degradation and word error rate (WER) increase due to DP to less than 2% and 7% respectively with 1 million clients and a stringent privacy budget of {\epsilon}=2.
PDF 16 pages, 10 figures, 5 tables
点此查看论文截图
Enhancing Language Models for Financial Relation Extraction with Named Entities and Part-of-Speech
Authors:Menglin Li, Kwan Hui Lim
The Financial Relation Extraction (FinRE) task involves identifying the entities and their relation, given a piece of financial statement/text. To solve this FinRE problem, we propose a simple but effective strategy that improves the performance of pre-trained language models by augmenting them with Named Entity Recognition (NER) and Part-Of-Speech (POS), as well as different approaches to combine these information. Experiments on a financial relations dataset show promising results and highlights the benefits of incorporating NER and POS in existing models. Our dataset and codes are available at https://github.com/kwanhui/FinRelExtract.
PDF Accepted to ICLR 2024 Tiny Paper Track
点此查看论文截图