Xie Chen

Chenpeng Du, Qi Chen, Tianyu He et al. · microsoft-research

While recent research has made significant progress in speech-driven talking face generation, the quality of the generated video still lags behind that of real recordings. One reason for this is the use of handcrafted intermediate representations like facial landmarks and 3DMM coefficients, which are designed based on human knowledge and are insufficient to precisely describe facial movements. Additionally, these methods require an external pretrained model for extracting these representations, whose performance sets an upper bound on talking face generation. To address these limitations, we propose a novel method called DAE-Talker that leverages data-driven latent representations obtained from a diffusion autoencoder (DAE). DAE contains an image encoder that encodes an image into a latent vector and a DDIM image decoder that reconstructs the image from it. We train our DAE on talking face video frames and then extract their latent representations as the training target for a Conformer-based speech2latent model. This allows DAE-Talker to synthesize full video frames and produce natural head movements that align with the content of speech, rather than relying on a predetermined head pose from a template video. We also introduce pose modelling in speech2latent for pose controllability. Additionally, we propose a novel method for generating continuous video frames with the DDIM image decoder trained on individual frames, eliminating the need for modelling the joint distribution of consecutive frames directly. Our experiments show that DAE-Talker outperforms existing popular methods in lip-sync, video fidelity, and pose naturalness. We also conduct ablation studies to analyze the effectiveness of the proposed techniques and demonstrate the pose controllability of DAE-Talker.

Yiwei Guo, Chenpeng Du, Ziyang Ma et al.

Although diffusion models in text-to-speech have become a popular choice due to their strong generative ability, the intrinsic complexity of sampling from diffusion models harms their efficiency. Alternatively, we propose VoiceFlow, an acoustic model that utilizes a rectified flow matching algorithm to achieve high synthesis quality with a limited number of sampling steps. VoiceFlow formulates the process of generating mel-spectrograms into an ordinary differential equation conditional on text inputs, whose vector field is then estimated. The rectified flow technique then effectively straightens its sampling trajectory for efficient synthesis. Subjective and objective evaluations on both single and multi-speaker corpora showed the superior synthesis quality of VoiceFlow compared to the diffusion counterpart. Ablation studies further verified the validity of the rectified flow technique in VoiceFlow.

Yiwei Guo, Chenpeng Du, Xie Chen et al.

Although current neural text-to-speech (TTS) models are able to generate high-quality speech, intensity controllable emotional TTS is still a challenging task. Most existing methods need external optimizations for intensity calculation, leading to suboptimal results or degraded quality. In this paper, we propose EmoDiff, a diffusion-based TTS model where emotion intensity can be manipulated by a proposed soft-label guidance technique derived from classifier guidance. Specifically, instead of being guided with a one-hot vector for the specified emotion, EmoDiff is guided with a soft label where the value of the specified emotion and \textit{Neutral} is set to $α$ and $1-α$ respectively. The $α$ here represents the emotion intensity and can be chosen from 0 to 1. Our experiments show that EmoDiff can precisely control the emotion intensity while maintaining high voice quality. Moreover, diverse speech with specified emotion intensity can be generated by sampling in the reverse denoising process.

Xun Gong, Yu Wu, Jinyu Li et al.

Traditional automatic speech recognition~(ASR) systems usually focus on individual utterances, without considering long-form speech with useful historical information, which is more practical in real scenarios. Simply attending longer transcription history for a vanilla neural transducer model shows no much gain in our preliminary experiments, since the prediction network is not a pure language model. This motivates us to leverage the factorized neural transducer structure, containing a real language model, the vocabulary predictor. We propose the {LongFNT-Text} architecture, which fuses the sentence-level long-form features directly with the output of the vocabulary predictor and then embeds token-level long-form features inside the vocabulary predictor, with a pre-trained contextual encoder RoBERTa to further boost the performance. Moreover, we propose the {LongFNT} architecture by extending the long-form speech to the original speech input and achieve the best performance. The effectiveness of our LongFNT approach is validated on LibriSpeech and GigaSpeech corpora with 19% and 12% relative word error rate~(WER) reduction, respectively.

Zheng Liang, Zheshu Song, Ziyang Ma et al.

Recently, end-to-end (E2E) automatic speech recognition (ASR) models have made great strides and exhibit excellent performance in general speech recognition. However, there remain several challenging scenarios that E2E models are not competent in, such as code-switching and named entity recognition (NER). Data augmentation is a common and effective practice for these two scenarios. However, the current data augmentation methods mainly rely on audio splicing and text-to-speech (TTS) models, which might result in discontinuous, unrealistic, and less diversified speech. To mitigate these potential issues, we propose a novel data augmentation method by applying the text-based speech editing model. The augmented speech from speech editing systems is more coherent and diversified, also more akin to real speech. The experimental results on code-switching and NER tasks show that our proposed method can significantly outperform the audio splicing and neural TTS based data augmentation systems.

Ziyang Ma, Zhisheng Zheng, Guanrou Yang et al.

The excellent generalization ability of self-supervised learning (SSL) for speech foundation models has garnered significant attention. HuBERT is a successful example that utilizes offline clustering to convert speech features into discrete units for a masked language modeling pretext task. However, simply clustering features as targets by k-means does not fully inspire the model's performance. In this work, we present an unsupervised method to improve SSL targets. Two models are proposed, MonoBERT and PolyBERT, which leverage context-independent and context-dependent phoneme-based units for pre-training. Our models outperform other SSL models significantly on the LibriSpeech benchmark without the need for iterative re-clustering and re-training. Furthermore, our models equipped with context-dependent units even outperform target-improvement models that use labeled data during pre-training. How we progressively improve the unit discovery process is demonstrated through experiments.

Ziyang Ma, Wen Wu, Zhisheng Zheng et al.

In this paper, we explored how to boost speech emotion recognition (SER) with the state-of-the-art speech pre-trained model (PTM), data2vec, text generation technique, GPT-4, and speech synthesis technique, Azure TTS. First, we investigated the representation ability of different speech self-supervised pre-trained models, and we found that data2vec has a good representation ability on the SER task. Second, we employed a powerful large language model (LLM), GPT-4, and emotional text-to-speech (TTS) model, Azure TTS, to generate emotionally congruent text and speech. We carefully designed the text prompt and dataset construction, to obtain the synthetic emotional speech data with high quality. Third, we studied different ways of data augmentation to promote the SER task with synthetic speech, including random mixing, adversarial training, transfer learning, and curriculum learning. Experiments and ablation studies on the IEMOCAP dataset demonstrate the effectiveness of our method, compared with other data augmentation methods, and data augmentation with other synthetic data.

Chen Xie, Xiaodong Gu, Yuling Shi et al.

Code complexity metrics such as cyclomatic complexity have long been used to assess software quality and maintainability. With the rapid advancement of large language models (LLMs) on coding tasks, an important yet underexplored question arises: do traditional complexity metrics meaningfully characterize the coding difficulty that LLMs perceive? In this work, we empirically demonstrate that classical complexity metrics exhibit no consistent correlation with LLM performance, revealing a fundamental mismatch with model-perceived difficulty. To address this gap, we propose LM-CC, a novel code complexity metric tailored for LLMs, grounded in the hypothesis that model-perceived code difficulty is fundamentally driven by semantic nonlinearity. LM-CC quantifies complexity through an entropy-guided semantic compositional hierarchy, capturing the cumulative uncertainty encountered by LLMs during code understanding. Our experimental results demonstrate that LM-CC exhibits strong and consistent partial correlations with LLM performance, while semantics-preserving reductions in LM-CC consistently lead to improved downstream task performance. The source code is available at: https://github.com/xchen121/lm-cc.

Junzhe Liu, Jianwei Yu, Xie Chen

Adapting End-to-End ASR models to out-of-domain datasets with text data is challenging. Factorized neural Transducer (FNT) aims to address this issue by introducing a separate vocabulary decoder to predict the vocabulary. Nonetheless, this approach has limitations in fusing acoustic and language information seamlessly. Moreover, a degradation in word error rate (WER) on the general test sets was also observed, leading to doubts about its overall performance. In response to this challenge, we present the improved factorized neural Transducer (IFNT) model structure designed to comprehensively integrate acoustic and language information while enabling effective text adaptation. We assess the performance of our proposed method on English and Mandarin datasets. The results indicate that IFNT not only surpasses the neural Transducer and FNT in baseline performance in both scenarios but also exhibits superior adaptation ability compared to FNT. On source domains, IFNT demonstrated statistically significant accuracy improvements, achieving a relative enhancement of 1.2% to 2.8% in baseline accuracy compared to the neural Transducer. On out-of-domain datasets, IFNT shows relative WER(CER) improvements of up to 30.2% over the standard neural Transducer with shallow fusion, and relative WER(CER) reductions ranging from 1.1% to 2.8% on test sets compared to the FNT model.

Ziyang Ma, Zhisheng Zheng, Changli Tang et al.

In this paper, we provide a new perspective on self-supervised speech models from how the training targets are obtained. We generalize the targets extractor into Offline Targets Extractor (Off-TE) and Online Targets Extractor (On-TE). Based on this, we propose a new multi-tasking learning framework for self-supervised learning, MT4SSL, which stands for Boosting Self-Supervised Speech Representation Learning by Integrating Multiple Targets. MT4SSL uses the K-means algorithm as an Off-TE and a teacher network without gradients as an On-TE, respectively. Our model outperforms previous SSL methods by nontrivial margins on the LibriSpeech benchmark, and is comparable to or even better than the best-performing models with fewer data. Furthermore, we find that using both Off-TE and On-TE results in better convergence in the pre-training phase. With both effectiveness and efficiency, we think doing multi-task learning on self-supervised speech models from our perspective is a promising trend.

Mingyu Cui, Jiawen Kang, Jiajun Deng et al.

Current ASR systems are mainly trained and evaluated at the utterance level. Long range cross utterance context can be incorporated. A key task is to derive a suitable compact representation of the most relevant history contexts. In contrast to previous researches based on either LSTM-RNN encoded histories that attenuate the information from longer range contexts, or frame level concatenation of transformer context embeddings, in this paper compact low-dimensional cross utterance contextual features are learned in the Conformer-Transducer Encoder using specially designed attention pooling layers that are applied over efficiently cached preceding utterances history vectors. Experiments on the 1000-hr Gigaspeech corpus demonstrate that the proposed contextualized streaming Conformer-Transducers outperform the baseline using utterance internal context only with statistically significant WER reductions of 0.7% to 0.5% absolute (4.3% to 3.1% relative) on the dev and test data.

Xie Chen, Ziyang Ma, Changli Tang et al.

Recent years have witnessed a boom in self-supervised learning (SSL) in various areas including speech processing. Speech based SSL models present promising performance in a range of speech related tasks. However, the training of SSL models is computationally expensive and a common practice is to fine-tune a released SSL model on the specific task. It is essential to use consistent front-end input during pre-training and fine-tuning. This consistency may introduce potential issues when the optimal front-end is not the same as that used in pre-training. In this paper, we propose a simple but effective front-end adapter to address this front-end discrepancy. By minimizing the distance between the outputs of different front-ends, the filterbank feature (Fbank) can be compatible with SSL models which are pre-trained with waveform. The experiment results demonstrate the effectiveness of our proposed front-end adapter on several popular SSL models for the speech recognition task.

Sen Liu, Yiwei Guo, Chenpeng Du et al.

Although high-fidelity speech can be obtained for intralingual speech synthesis, cross-lingual text-to-speech (CTTS) is still far from satisfactory as it is difficult to accurately retain the speaker timbres(i.e. speaker similarity) and eliminate the accents from their first language(i.e. nativeness). In this paper, we demonstrated that vector-quantized(VQ) acoustic feature contains less speaker information than mel-spectrogram. Based on this finding, we propose a novel dual speaker embedding TTS (DSE-TTS) framework for CTTS with authentic speaking style. Here, one embedding is fed to the acoustic model to learn the linguistic speaking style, while the other one is integrated into the vocoder to mimic the target speaker's timbre. Experiments show that by combining both embeddings, DSE-TTS significantly outperforms the state-of-the-art SANE-TTS in cross-lingual synthesis, especially in terms of nativeness.

Yakun Song, Zhuo Chen, Xiaofei Wang et al.

The language model (LM) approach based on acoustic and linguistic prompts, such as VALL-E, has achieved remarkable progress in the field of zero-shot audio generation. However, existing methods still have some limitations: 1) repetitions, transpositions, and omissions in the output synthesized speech due to limited alignment constraints between audio and phoneme tokens; 2) challenges of fine-grained control over the synthesized speech with autoregressive (AR) language model; 3) infinite silence generation due to the nature of AR-based decoding, especially under the greedy strategy. To alleviate these issues, we propose ELLA-V, a simple but efficient LM-based zero-shot text-to-speech (TTS) framework, which enables fine-grained control over synthesized audio at the phoneme level. The key to ELLA-V is interleaving sequences of acoustic and phoneme tokens, where phoneme tokens appear ahead of the corresponding acoustic tokens. The experimental findings reveal that our model outperforms VALL-E in terms of accuracy and delivers more stable results using both greedy and sampling-based decoding strategies. The code of ELLA-V will be open-sourced after cleanups. Audio samples are available at https://ereboas.github.io/ELLAV/.

Peng Wang, Yifan Yang, Zheng Liang et al.

Despite advancements of end-to-end (E2E) models in speech recognition, named entity recognition (NER) is still challenging but critical for semantic understanding. Previous studies mainly focus on various rule-based or attention-based contextual biasing algorithms. However, their performance might be sensitive to the biasing weight or degraded by excessive attention to the named entity list, along with a risk of false triggering. Inspired by the success of the class-based language model (LM) in NER in conventional hybrid systems and the effective decoupling of acoustic and linguistic information in the factorized neural Transducer (FNT), we propose C-FNT, a novel E2E model that incorporates class-based LMs into FNT. In C-FNT, the LM score of named entities can be associated with the name class instead of its surface form. The experimental results show that our proposed C-FNT significantly reduces error in named entities without hurting performance in general word recognition.

Hanglei Zhang, Yiwei Guo, Sen Liu et al.

Expressive text-to-speech (TTS) aims to synthesize speeches with human-like tones, moods, or even artistic attributes. Recent advancements in expressive TTS empower users with the ability to directly control synthesis style through natural language prompts. However, these methods often require excessive training with a significant amount of style-annotated data, which can be challenging to acquire. Moreover, they may have limited adaptability due to fixed style annotations. In this work, we present FreeStyleTTS (FS-TTS), a controllable expressive TTS model with minimal human annotations. Our approach utilizes a large language model (LLM) to transform expressive TTS into a style retrieval task. The LLM selects the best-matching style references from annotated utterances based on external style prompts, which can be raw input text or natural language style descriptions. The selected reference guides the TTS pipeline to synthesize speeches with the intended style. This innovative approach provides flexible, versatile, and precise style control with minimal human workload. Experiments on a Mandarin storytelling corpus demonstrate FS-TTS's proficiency in leveraging LLM's semantic inference ability to retrieve desired styles from either input text or user-defined descriptions. This results in synthetic speeches that are closely aligned with the specified styles.

Pingyi Fan, Anbai Jiang, Shuwei Zhang et al.

With the rapid deployment of SCADA systems, how to effectively analyze industrial signals and detect abnormal states is an urgent need for the industry. Due to the significant heterogeneity of these signals, which we summarize as the M5 problem, previous works only focus on small sub-problems and employ specialized models, failing to utilize the synergies between modalities and the powerful scaling law. However, we argue that the M5 signals can be modeled in a unified manner due to the intrinsic similarity. As a result, we propose FISHER, a Foundation model for multi-modal Industrial Signal compreHEnsive Representation. To support arbitrary sampling rates, FISHER considers the increment of sampling rate as the concatenation of sub-band information. Specifically, FISHER takes the STFT sub-band as the modeling unit and adopts a teacher student SSL framework for pre-training. We also develop the RMIS benchmark, which evaluates the representations of M5 industrial signals on multiple health management tasks. Compared with top SSL models, FISHER showcases versatile and outstanding capabilities with a general performance gain up to 5.03%, along with much more efficient scaling curves. We also investigate the scaling law on downstream tasks and derive potential avenues for future works. FISHER is now open-sourced on https://github.com/jianganbai/FISHER

Yifan Yang, Xiaoyu Yang, Liyong Guo et al.

Neural Transducer and connectionist temporal classification (CTC) are popular end-to-end automatic speech recognition systems. Due to their frame-synchronous design, blank symbols are introduced to address the length mismatch between acoustic frames and output tokens, which might bring redundant computation. Previous studies managed to accelerate the training and inference of neural Transducers by discarding frames based on the blank symbols predicted by a co-trained CTC. However, there is no guarantee that the co-trained CTC can maximize the ratio of blank symbols. This paper proposes two novel regularization methods to explicitly encourage more blanks by constraining the self-loop of non-blank symbols in the CTC. It is interesting to find that the frame reduction ratio of the neural Transducer can approach the theoretical boundary. Experiments on LibriSpeech corpus show that our proposed method accelerates the inference of neural Transducer by 4 times without sacrificing performance. Our work is open-sourced and publicly available https://github.com/k2-fsa/icefall.

Wenxi Chen, Yuzhe Liang, Ziyang Ma et al.

Audio self-supervised learning (SSL) pre-training, which aims to learn good representations from unlabeled audio, has made remarkable progress. However, the extensive computational demands during pre-training pose a significant barrier to the potential application and optimization of audio SSL models. In this paper, inspired by the success of data2vec 2.0 in image modality and Audio-MAE in audio modality, we introduce Efficient Audio Transformer (EAT) to further improve the effectiveness and efficiency in audio SSL. The proposed EAT adopts the bootstrap self-supervised training paradigm to the audio domain. A novel Utterance-Frame Objective (UFO) is designed to enhance the modeling capability of acoustic events. Furthermore, we reveal that the masking strategy is critical in audio SSL pre-training, and superior audio representations can be obtained with large inverse block masks. Experiment results demonstrate that EAT achieves state-of-the-art (SOTA) performance on a range of audio-related tasks, including AudioSet (AS-2M, AS-20K), ESC-50, and SPC-2, along with a significant pre-training speedup up to ~15x compared to existing audio SSL models.

Junjie Li, Yiwei Guo, Xie Chen et al.

Zero-shot voice conversion (VC) aims to transfer the source speaker timbre to arbitrary unseen target speaker timbre, while keeping the linguistic content unchanged. Although the voice of generated speech can be controlled by providing the speaker embedding of the target speaker, the speaker similarity still lags behind the ground truth recordings. In this paper, we propose SEF-VC, a speaker embedding free voice conversion model, which is designed to learn and incorporate speaker timbre from reference speech via a powerful position-agnostic cross-attention mechanism, and then reconstruct waveform from HuBERT semantic tokens in a non-autoregressive manner. The concise design of SEF-VC enhances its training stability and voice conversion performance. Objective and subjective evaluations demonstrate the superiority of SEF-VC to generate high-quality speech with better similarity to target reference than strong zero-shot VC baselines, even for very short reference speeches.

Yiwei Guo, Zhihan Li, Chenpeng Du et al.

Although discrete speech tokens have exhibited strong potential for language model-based speech generation, their high bitrates and redundant timbre information restrict the development of such models. In this work, we propose LSCodec, a discrete speech codec that has both low bitrate and speaker decoupling ability. LSCodec adopts a multi-stage unsupervised training framework with a speaker perturbation technique. A continuous information bottleneck is first established, followed by vector quantization that produces a discrete speaker-decoupled space. A discrete token vocoder finally refines acoustic details from LSCodec. By reconstruction evaluations, LSCodec demonstrates superior intelligibility and audio quality with only a single codebook and smaller vocabulary size than baselines. Voice conversion and speaker probing experiments prove the excellent speaker disentanglement of LSCodec, and ablation study verifies the effectiveness of the proposed training framework.

Yifan Yang, Shujie Liu, Jinyu Li et al.

Recent zero-shot text-to-speech (TTS) systems face a common dilemma: autoregressive (AR) models suffer from slow generation and lack duration controllability, while non-autoregressive (NAR) models lack temporal modeling and typically require complex designs. In this paper, we introduce a novel pseudo-autoregressive (PAR) codec language modeling approach that unifies AR and NAR modeling. Combining explicit temporal modeling from AR with parallel generation from NAR, PAR generates dynamic-length spans at fixed time steps. Building on PAR, we propose PALLE, a two-stage TTS system that leverages PAR for initial generation followed by NAR refinement. In the first stage, PAR progressively generates speech tokens along the time dimension, with each step predicting all positions in parallel but only retaining the left-most span. In the second stage, low-confidence tokens are iteratively refined in parallel, leveraging the global contextual information. Experiments demonstrate that PALLE, trained on LibriTTS, outperforms state-of-the-art systems trained on large-scale data, including F5-TTS, E2-TTS, and MaskGCT, on the LibriSpeech test-clean set in terms of speech quality, speaker similarity, and intelligibility, while achieving up to ten times faster inference speed. Audio samples are available at https://microsoft.com/research/project/vall-e-x/palle.

Hanglei Zhang, Yiwei Guo, Zhihan Li et al.

Most neural speech codecs achieve bitrate adjustment through intra-frame mechanisms, such as codebook dropout, at a Constant Frame Rate (CFR). However, speech segments inherently have time-varying information density (e.g., silent intervals versus voiced regions). This property makes CFR not optimal in terms of bitrate and token sequence length, hindering efficiency in real-time applications. In this work, we propose a Temporally Flexible Coding (TFC) technique, introducing variable frame rate (VFR) into neural speech codecs for the first time. TFC enables seamlessly tunable average frame rates and dynamically allocates frame rates based on temporal entropy. Experimental results show that a codec with TFC achieves optimal reconstruction quality with high flexibility, and maintains competitive performance even at lower frame rates. Our approach is promising for the integration with other efforts to develop low-frame-rate neural speech codecs for more efficient downstream tasks.

Xiquan Li, Junxi Liu, Yuzhe Liang et al.

Recent years have witnessed remarkable progress in Text-to-Audio Generation (TTA), providing sound creators with powerful tools to transform inspirations into vivid audio. Yet despite these advances, current TTA systems often suffer from slow inference speed, which greatly hinders the efficiency and smoothness of audio creation. In this paper, we present MeanAudio, a fast and faithful text-to-audio generator capable of rendering realistic sound with only one function evaluation (1-NFE). MeanAudio leverages: (i) the MeanFlow objective with guided velocity target that significantly accelerates inference speed, (ii) an enhanced Flux-style transformer with dual text encoders for better semantic alignment and synthesis quality, and (iii) an efficient instantaneous-to-mean curriculum that speeds up convergence and enables training on consumer-grade GPUs. Through a comprehensive evaluation study, we demonstrate that MeanAudio achieves state-of-the-art performance in single-step audio generation. Specifically, it achieves a real-time factor (RTF) of 0.013 on a single NVIDIA RTX 3090, yielding a 100x speedup over SOTA diffusion-based TTA systems. Moreover, MeanAudio also shows strong performance in multi-step generation, enabling smooth transitions across successive synthesis steps.

Keqi Deng, Wenxi Chen, Xie Chen et al.

Simultaneous speech translation (SST) outputs translations in parallel with streaming speech input, balancing translation quality and latency. While large language models (LLMs) have been extended to handle the speech modality, streaming remains challenging as speech is prepended as a prompt for the entire generation process. To unlock LLM streaming capability, this paper proposes SimulS2S-LLM, which trains speech LLMs offline and employs a test-time policy to guide simultaneous inference. SimulS2S-LLM alleviates the mismatch between training and inference by extracting boundary-aware speech prompts that allows it to be better matched with text input data. SimulS2S-LLM achieves simultaneous speech-to-speech translation (Simul-S2ST) by predicting discrete output speech tokens and then synthesising output speech using a pre-trained vocoder. An incremental beam search is designed to expand the search space of speech token prediction without increasing latency. Experiments on the CVSS speech data show that SimulS2S-LLM offers a better translation quality-latency trade-off than existing methods that use the same training data, such as improving ASR-BLEU scores by 3 points at similar latency.

Yiwei Guo, Chenrun Wang, Yifan Yang et al.

Discrete speech tokens have been more and more popular in multiple speech processing fields, including automatic speech recognition (ASR), text-to-speech (TTS) and singing voice synthesis (SVS). In this paper, we describe the systems developed by the SJTU X-LANCE group for the TTS (acoustic + vocoder), SVS, and ASR tracks in the Interspeech 2024 Speech Processing Using Discrete Speech Unit Challenge. Notably, we achieved 1st rank on the leaderboard in the TTS track both with the whole training set and only 1h training data, with the highest UTMOS score and lowest bitrate among all submissions.

Bo Chen, Tao Liu, Qi Chen et al.

Talking face generation aims to synthesize realistic speaking portraits from a single image, yet existing methods often rely on explicit optical flow and local warping, which fail to model complex global motions and cause identity drift. We present IMTalker, a novel framework that achieves efficient and high-fidelity talking face generation through implicit motion transfer. The core idea is to replace traditional flow-based warping with a cross-attention mechanism that implicitly models motion discrepancy and identity alignment within a unified latent space, enabling robust global motion rendering. To further preserve speaker identity during cross-identity reenactment, we introduce an identity-adaptive module that projects motion latents into personalized spaces, ensuring clear disentanglement between motion and identity. In addition, a lightweight flow-matching motion generator produces vivid and controllable implicit motion vectors from audio, pose, and gaze cues. Extensive experiments demonstrate that IMTalker surpasses prior methods in motion accuracy, identity preservation, and audio-lip synchronization, achieving state-of-the-art quality with superior efficiency, operating at 40 FPS for video-driven and 42 FPS for audio-driven generation on an RTX 4090 GPU. We will release our code and pre-trained models to facilitate applications and future research.

Junhao Xu, Xie Chen, Shoukang Hu et al.

The high memory consumption and computational costs of Recurrent neural network language models (RNNLMs) limit their wider application on resource constrained devices. In recent years, neural network quantization techniques that are capable of producing extremely low-bit compression, for example, binarized RNNLMs, are gaining increasing research interests. Directly training of quantized neural networks is difficult. By formulating quantized RNNLMs training as an optimization problem, this paper presents a novel method to train quantized RNNLMs from scratch using alternating direction methods of multipliers (ADMM). This method can also flexibly adjust the trade-off between the compression rate and model performance using tied low-bit quantization tables. Experiments on two tasks: Penn Treebank (PTB), and Switchboard (SWBD) suggest the proposed ADMM quantization achieved a model size compression factor of up to 31 times over the full precision baseline RNNLMs. Faster convergence of 5 times in model training over the baseline binarized RNNLM quantization was also obtained. Index Terms: Language models, Recurrent neural networks, Quantization, Alternating direction methods of multipliers.

Zhong Meng, Yashesh Gaur, Naoyuki Kanda et al.

Text-only adaptation of an end-to-end (E2E) model remains a challenging task for automatic speech recognition (ASR). Language model (LM) fusion-based approaches require an additional external LM during inference, significantly increasing the computation cost. To overcome this, we propose an internal LM adaptation (ILMA) of the E2E model using text-only data. Trained with audio-transcript pairs, an E2E model implicitly learns an internal LM that characterizes the token sequence probability which is approximated by the E2E model output after zeroing out the encoder contribution. During ILMA, we fine-tune the internal LM, i.e., the E2E components excluding the encoder, to minimize a cross-entropy loss. To make ILMA effective, it is essential to train the E2E model with an internal LM loss besides the standard E2E loss. Furthermore, we propose to regularize ILMA by minimizing the Kullback-Leibler divergence between the output distributions of the adapted and unadapted internal LMs. ILMA is the most effective when we update only the last linear layer of the joint network. ILMA enables a fast text-only adaptation of the E2E model without increasing the run-time computational cost. Experimented with 30K-hour trained transformer transducer models, ILMA achieves up to 34.9% relative word error rate reduction from the unadapted baseline.

Xie Chen, Zhong Meng, Sarangarajan Parthasarathy et al.

In recent years, end-to-end (E2E) based automatic speech recognition (ASR) systems have achieved great success due to their simplicity and promising performance. Neural Transducer based models are increasingly popular in streaming E2E based ASR systems and have been reported to outperform the traditional hybrid system in some scenarios. However, the joint optimization of acoustic model, lexicon and language model in neural Transducer also brings about challenges to utilize pure text for language model adaptation. This drawback might prevent their potential applications in practice. In order to address this issue, in this paper, we propose a novel model, factorized neural Transducer, by factorizing the blank and vocabulary prediction, and adopting a standalone language model for the vocabulary prediction. It is expected that this factorization can transfer the improvement of the standalone language model to the Transducer for speech recognition, which allows various language model adaptation techniques to be applied. We demonstrate that the proposed factorized neural Transducer yields 15% to 20% WER improvements when out-of-domain text data is used for language model adaptation, at the cost of a minor degradation in WER on a general test set.

Zhong Meng, Naoyuki Kanda, Yashesh Gaur et al.

The efficacy of external language model (LM) integration with existing end-to-end (E2E) automatic speech recognition (ASR) systems can be improved significantly using the internal language model estimation (ILME) method. In this method, the internal LM score is subtracted from the score obtained by interpolating the E2E score with the external LM score, during inference. To improve the ILME-based inference, we propose an internal LM training (ILMT) method to minimize an additional internal LM loss by updating only the E2E model components that affect the internal LM estimation. ILMT encourages the E2E model to form a standalone LM inside its existing components, without sacrificing ASR accuracy. After ILMT, the more modular E2E model with matched training and inference criteria enables a more thorough elimination of the source-domain internal LM, and therefore leads to a more effective integration of the target-domain external LM. Experimented with 30K-hour trained recurrent neural network transducer and attention-based encoder-decoder models, ILMT with ILME-based inference achieves up to 31.5% and 11.4% relative word error rate reductions from standard E2E training with Shallow Fusion on out-of-domain LibriSpeech and in-domain Microsoft production test sets, respectively.

Xie Chen, Yu Wu, Zhenghao Wang et al.

Recently, Transformer based end-to-end models have achieved great success in many areas including speech recognition. However, compared to LSTM models, the heavy computational cost of the Transformer during inference is a key issue to prevent their applications. In this work, we explored the potential of Transformer Transducer (T-T) models for the fist pass decoding with low latency and fast speed on a large-scale dataset. We combine the idea of Transformer-XL and chunk-wise streaming processing to design a streamable Transformer Transducer model. We demonstrate that T-T outperforms the hybrid model, RNN Transducer (RNN-T), and streamable Transformer attention-based encoder-decoder model in the streaming scenario. Furthermore, the runtime cost and latency can be optimized with a relatively small look-ahead.

Xie Chen, Xunying Liu, Anton Ragni et al.

Recently, bidirectional recurrent network language models (bi-RNNLMs) have been shown to outperform standard, unidirectional, recurrent neural network language models (uni-RNNLMs) on a range of speech recognition tasks. This indicates that future word context information beyond the word history can be useful. However, bi-RNNLMs pose a number of challenges as they make use of the complete previous and future word context information. This impacts both training efficiency and their use within a lattice rescoring framework. In this paper these issues are addressed by proposing a novel neural network structure, succeeding word RNNLMs (su-RNNLMs). Instead of using a recurrent unit to capture the complete future word contexts, a feedforward unit is used to model a finite number of succeeding, future, words. This model can be trained much more efficiently than bi-RNNLMs and can also be used for lattice rescoring. Experimental results on a meeting transcription task (AMI) show the proposed model consistently outperformed uni-RNNLMs and yield only a slight degradation compared to bi-RNNLMs in N-best rescoring. Additionally, performance improvements can be obtained using lattice rescoring and subsequent confusion network decoding.