Speech-driven 3D facial animation synthesis has been a challenging task both in industry and research. Recent methods mostly focus on deterministic deep learning methods meaning that given a speech input, the output is always the same. However, in reality, the non-verbal facial cues that reside throughout the face are non-deterministic in nature. In addition, majority of the approaches focus on 3D vertex based datasets and methods that are compatible with existing facial animation pipelines with rigged characters is scarce. To eliminate these issues, we present FaceDiffuser, a non-deterministic deep learning model to generate speech-driven facial animations that is trained with both 3D vertex and blendshape based datasets. Our method is based on the diffusion technique and uses the pre-trained large speech representation model HuBERT to encode the audio input. To the best of our knowledge, we are the first to employ the diffusion method for the task of speech-driven 3D facial animation synthesis. We have run extensive objective and subjective analyses and show that our approach achieves better or comparable results in comparison to the state-of-the-art methods. We also introduce a new in-house dataset that is based on a blendshape based rigged character. We recommend watching the accompanying supplementary video. The code and the dataset will be publicly available.
The Conformer has become the most popular encoder model for automatic speech recognition (ASR). It adds convolution modules to a transformer to learn both local and global dependencies. In this work we describe a faster, more memory-efficient, and better-performing transformer, called Zipformer. Modeling changes include: 1) a U-Net-like encoder structure where middle stacks operate at lower frame rates; 2) reorganized block structure with more modules, within which we re-use attention weights for efficiency; 3) a modified form of LayerNorm called BiasNorm allows us to retain some length information; 4) new activation functions SwooshR and SwooshL work better than Swish. We also propose a new optimizer, called ScaledAdam, which scales the update by each tensor's current scale to keep the relative change about the same, and also explictly learns the parameter scale. It achieves faster convergence and better performance than Adam. Extensive experiments on LibriSpeech, Aishell-1, and WenetSpeech datasets demonstrate the effectiveness of our proposed Zipformer over other state-of-the-art ASR models. Our code is publicly available at https://github.com/k2-fsa/icefall.
Current speaker anonymization methods, especially with self-supervised learning (SSL) models, require massive computational resources when hiding speaker identity. This paper proposes an effective and parameter-efficient speaker anonymization method based on recent End-to-End model reprogramming technology. To improve the anonymization performance, we first extract speaker representation from large SSL models as the speaker identifies. To hide the speaker's identity, we reprogram the speaker representation by adapting the speaker to a pseudo domain. Extensive experiments are carried out on the VoicePrivacy Challenge (VPC) 2022 datasets to demonstrate the effectiveness of our proposed parameter-efficient learning anonymization methods. Additionally, while achieving comparable performance with the VPC 2022 strong baseline 1.b, our approach consumes less computational resources during anonymization.
We describe an end-to-end speech synthesis system that uses generative adversarial training. We train our Vocoder for raw phoneme-to-audio conversion, using explicit phonetic, pitch and duration modeling. We experiment with several pre-trained models for contextualized and decontextualized word embeddings and we introduce a new method for highly expressive character voice matching, based on discreet style tokens.
In this paper, we propose a new unsupervised domain adaptation (DA) method called layer-adapted implicit distribution alignment networks (LIDAN) to address the challenge of cross-corpus speech emotion recognition (SER). LIDAN extends our previous ICASSP work, deep implicit distribution alignment networks (DIDAN), whose key contribution lies in the introduction of a novel regularization term called implicit distribution alignment (IDA). This term allows DIDAN trained on source (training) speech samples to remain applicable to predicting emotion labels for target (testing) speech samples, regardless of corpus variance in cross-corpus SER. To further enhance this method, we extend IDA to layer-adapted IDA (LIDA), resulting in LIDAN. This layer-adpated extention consists of three modified IDA terms that consider emotion labels at different levels of granularity. These terms are strategically arranged within different fully connected layers in LIDAN, aligning with the increasing emotion-discriminative abilities with respect to the layer depth. This arrangement enables LIDAN to more effectively learn emotion-discriminative and corpus-invariant features for SER across various corpora compared to DIDAN. It is also worthy to mention that unlike most existing methods that rely on estimating statistical moments to describe pre-assumed explicit distributions, both IDA and LIDA take a different approach. They utilize an idea of target sample reconstruction to directly bridge the feature distribution gap without making assumptions about their distribution type. As a result, DIDAN and LIDAN can be viewed as implicit cross-corpus SER methods. To evaluate LIDAN, we conducted extensive cross-corpus SER experiments on EmoDB, eNTERFACE, and CASIA corpora. The experimental results demonstrate that LIDAN surpasses recent state-of-the-art explicit unsupervised DA methods in tackling cross-corpus SER tasks.
Accurate spelling correction is a critical step in modern search interfaces, especially in an era of mobile devices and speech-to-text interfaces. For services that are deployed around the world, this poses a significant challenge for multilingual NLP: spelling errors need to be caught and corrected in all languages, and even in queries that use multiple languages. In this paper, we tackle this challenge using multi-teacher distillation. On our approach, a monolingual teacher model is trained for each language/locale, and these individual models are distilled into a single multilingual student model intended to serve all languages/locales. In experiments using open-source data as well as user data from a worldwide search service, we show that this leads to highly effective spelling correction models that can meet the tight latency requirements of deployed services.
Computational complexity is critical when deploying deep learning-based speech denoising models for on-device applications. Most prior research focused on optimizing model architectures to meet specific computational cost constraints, often creating distinct neural network architectures for different complexity limitations. This study conducts complexity scaling for speech denoising tasks, aiming to consolidate models with various complexities into a unified architecture. We present a Multi-Path Transform-based (MPT) architecture to handle both low- and high-complexity scenarios. A series of MPT networks present high performance covering a wide range of computational complexities on the DNS challenge dataset. Moreover, inspired by the scaling experiments in natural language processing, we explore the empirical relationship between model performance and computational cost on the denoising task. As the complexity number of multiply-accumulate operations (MACs) is scaled from 50M/s to 15G/s on MPT networks, we observe a linear increase in the values of PESQ-WB and SI-SNR, proportional to the logarithm of MACs, which might contribute to the understanding and application of complexity scaling in speech denoising tasks.
Soft random sampling (SRS) is a simple yet effective approach for efficient training of large-scale deep neural networks when dealing with massive data. SRS selects a subset uniformly at random with replacement from the full data set in each epoch. In this paper, we conduct a theoretical and empirical analysis of SRS. First, we analyze its sampling dynamics including data coverage and occupancy. Next, we investigate its convergence with non-convex objective functions and give the convergence rate. Finally, we provide its generalization performance. We empirically evaluate SRS for image recognition on CIFAR10 and automatic speech recognition on Librispeech and an in-house payload dataset to demonstrate its effectiveness. Compared to existing coreset-based data selection methods, SRS offers a better accuracy-efficiency trade-off. Especially on real-world industrial scale data sets, it is shown to be a powerful training strategy with significant speedup and competitive performance with almost no additional computing cost.
The Counter Narrative (CN) is a promising approach to combat online hate speech (HS) without infringing on freedom of speech. In recent years, there has been a growing interest in automatically generating CNs using natural language generation techniques. However, current automatic CN generation methods mainly rely on expert-authored datasets for training, which are time-consuming and labor-intensive to acquire. Furthermore, these methods cannot directly obtain and extend counter-knowledge from external statistics, facts, or examples. To address these limitations, we propose Retrieval-Augmented Unsupervised Counter Narrative Generation (RAUCG) to automatically expand external counter-knowledge and map it into CNs in an unsupervised paradigm. Specifically, we first introduce an SSF retrieval method to retrieve counter-knowledge from the multiple perspectives of stance consistency, semantic overlap rate, and fitness for HS. Then we design an energy-based decoding mechanism by quantizing knowledge injection, countering and fluency constraints into differentiable functions, to enable the model to build mappings from counter-knowledge to CNs without expert-authored CN data. Lastly, we comprehensively evaluate model performance in terms of language quality, toxicity, persuasiveness, relevance, and success rate of countering HS, etc. Experimental results show that RAUCG outperforms strong baselines on all metrics and exhibits stronger generalization capabilities, achieving significant improvements of +2.0% in relevance and +4.5% in success rate of countering metrics. Moreover, RAUCG enabled GPT2 to outperform T0 in all metrics, despite the latter being approximately eight times larger than the former. Warning: This paper may contain offensive or upsetting content!
Arabic is a complex language with many varieties and dialects spoken by over 450 millions all around the world. Due to the linguistic diversity and variations, it is challenging to build a robust and generalized ASR system for Arabic. In this work, we address this gap by developing and demoing a system, dubbed VoxArabica, for dialect identification (DID) as well as automatic speech recognition (ASR) of Arabic. We train a wide range of models such as HuBERT (DID), Whisper, and XLS-R (ASR) in a supervised setting for Arabic DID and ASR tasks. Our DID models are trained to identify 17 different dialects in addition to MSA. We finetune our ASR models on MSA, Egyptian, Moroccan, and mixed data. Additionally, for the remaining dialects in ASR, we provide the option to choose various models such as Whisper and MMS in a zero-shot setting. We integrate these models into a single web interface with diverse features such as audio recording, file upload, model selection, and the option to raise flags for incorrect outputs. Overall, we believe VoxArabica will be useful for a wide range of audiences concerned with Arabic research. Our system is currently running at https://cdce-206-12-100-168.ngrok.io/.