Dynamic Language Group-Based MoE: Enhancing Efficiency and Flexibility for Code-Switching Speech Recognition

The Mixture of Experts (MoE) approach is ideally suited for tackling multilingual and code-switching (CS) challenges due to its multi-expert architecture. This work introduces the DLG-MoE, which is optimized for bilingual and CS scenarios. Our novel Dynamic Language Group-based MoE layer features a language router with shared weights for explicit language modeling, while independent unsupervised routers within the language group handle attributes beyond language. This structure not only enhances expert extension capabilities but also supports dynamic top-k training, allowing for flexible inference across various top-k values and improving overall performance. The model requires no pre-training and supports streaming recognition, achieving state-of-the-art (SOTA) results with unmatched flexibility compared to other methods. The Code will be released.

LAFMA: A Latent Flow Matching Model for Text-to-Audio Generation

Recently, the application of diffusion models has facilitated the significant development of speech and audio generation. Nevertheless, the quality of samples generated by diffusion models still needs improvement. And the effectiveness of the method is accompanied by the extensive number of sampling steps, leading to an extended synthesis time necessary for generating high-quality audio. Previous Text-to-Audio (TTA) methods mostly used diffusion models in the latent space for audio generation. In this paper, we explore the integration of the Flow Matching (FM) model into the audio latent space for audio generation. The FM is an alternative simulation-free method that trains continuous normalization flows (CNF) based on regressing vector fields. We demonstrate that our model significantly enhances the quality of generated audio samples, achieving better performance than prior models. Moreover, it reduces the number of inference steps to ten steps almost without sacrificing performance.

MM-TTS: Multi-modal Prompt based Style Transfer for Expressive Text-to-Speech Synthesis

The style transfer task in Text-to-Speech refers to the process of transferring style information into text content to generate corresponding speech with a specific style. However, most existing style transfer approaches are either based on fixed emotional labels or reference speech clips, which cannot achieve flexible style transfer. Recently, some methods have adopted text descriptions to guide style transfer. In this paper, we propose a more flexible multi-modal and style controllable TTS framework named MM-TTS. It can utilize any modality as the prompt in unified multi-modal prompt space, including reference speech, emotional facial images, and text descriptions, to control the style of the generated speech in a system. The challenges of modeling such a multi-modal style controllable TTS mainly lie in two aspects:1)aligning the multi-modal information into a unified style space to enable the input of arbitrary modality as the style prompt in a single system, and 2)efficiently transferring the unified style representation into the given text content, thereby empowering the ability to generate prompt style-related voice. To address these problems, we propose an aligned multi-modal prompt encoder that embeds different modalities into a unified style space, supporting style transfer for different modalities. Additionally, we present a new adaptive style transfer method named Style Adaptive Convolutions to achieve a better style representation. Furthermore, we design a Rectified Flow based Refiner to solve the problem of over-smoothing Mel-spectrogram and generate audio of higher fidelity. Since there is no public dataset for multi-modal TTS, we construct a dataset named MEAD-TTS, which is related to the field of expressive talking head. Our experiments on the MEAD-TTS dataset and out-of-domain datasets demonstrate that MM-TTS can achieve satisfactory results based on multi-modal prompts.

ReFlow-TTS: A Rectified Flow Model for High-fidelity Text-to-Speech

The diffusion models including Denoising Diffusion Probabilistic Models (DDPM) and score-based generative models have demonstrated excellent performance in speech synthesis tasks. However, its effectiveness comes at the cost of numerous sampling steps, resulting in prolonged sampling time required to synthesize high-quality speech. This drawback hinders its practical applicability in real-world scenarios. In this paper, we introduce ReFlow-TTS, a novel rectified flow based method for speech synthesis with high-fidelity. Specifically, our ReFlow-TTS is simply an Ordinary Differential Equation (ODE) model that transports Gaussian distribution to the ground-truth Mel-spectrogram distribution by straight line paths as much as possible. Furthermore, our proposed approach enables high-quality speech synthesis with a single sampling step and eliminates the need for training a teacher model. Our experiments on LJSpeech Dataset show that our ReFlow-TTS method achieves the best performance compared with other diffusion based models. And the ReFlow-TTS with one step sampling achieves competitive performance compared with existing one-step TTS models.

Community Detection Graph Convolutional Network for Overlap-Aware Speaker Diarization

The clustering algorithm plays a crucial role in speaker diarization systems. However, traditional clustering algorithms suffer from the complex distribution of speaker embeddings and lack of digging potential relationships between speakers in a session. We propose a novel graph-based clustering approach called Community Detection Graph Convolutional Network (CDGCN) to improve the performance of the speaker diarization system. The CDGCN-based clustering method consists of graph generation, sub-graph detection, and Graph-based Overlapped Speech Detection (Graph-OSD). Firstly, the graph generation refines the local linkages among speech segments. Secondly the sub-graph detection finds the optimal global partition of the speaker graph. Finally, we view speaker clustering for overlap-aware speaker diarization as an overlapped community detection task and design a Graph-OSD component to output overlap-aware labels. By capturing local and global information, the speaker diarization system with CDGCN clustering outperforms the traditional Clustering-based Speaker Diarization (CSD) systems on the DIHARD III corpus.

Interpretable Style Transfer for Text-to-Speech with ControlVAE and Diffusion Bridge

With the demand for autonomous control and personalized speech generation, the style control and transfer in Text-to-Speech (TTS) is becoming more and more important. In this paper, we propose a new TTS system that can perform style transfer with interpretability and high fidelity. Firstly, we design a TTS system that combines variational autoencoder (VAE) and diffusion refiner to get refined mel-spectrograms. Specifically, a two-stage and a one-stage system are designed respectively, to improve the audio quality and the performance of style transfer. Secondly, a diffusion bridge of quantized VAE is designed to efficiently learn complex discrete style representations and improve the performance of style transfer. To have a better ability of style transfer, we introduce ControlVAE to improve the reconstruction quality and have good interpretability simultaneously. Experiments on LibriTTS dataset demonstrate that our method is more effective than baseline models.

Towards A Unified Conformer Structure: from ASR to ASV Task

Transformer has achieved extraordinary performance in Natural Language Processing and Computer Vision tasks thanks to its powerful self-attention mechanism, and its variant Conformer has become a state-of-the-art architecture in the field of Automatic Speech Recognition (ASR). However, the main-stream architecture for Automatic Speaker Verification (ASV) is convolutional Neural Networks, and there is still much room for research on the Conformer based ASV. In this paper, firstly, we modify the Conformer architecture from ASR to ASV with very minor changes. Length-Scaled Attention (LSA) method and Sharpness-Aware Minimizationis (SAM) are adopted to improve model generalization. Experiments conducted on VoxCeleb and CN-Celeb show that our Conformer based ASV achieves competitive performance compared with the popular ECAPA-TDNN. Secondly, inspired by the transfer learning strategy, ASV Conformer is natural to be initialized from the pretrained ASR model. Via parameter transferring, self-attention mechanism could better focus on the relationship between sequence features, brings about 11% relative improvement in EER on test set of VoxCeleb and CN-Celeb, which reveals the potential of Conformer to unify ASV and ASR task. Finally, we provide a runtime in ASV-Subtools to evaluate its inference speed in production scenario. Our code is released at https://github.com/Snowdar/asv-subtools/tree/master/doc/papers/conformer.md.

Spatial-aware Speaker Diarization for Multi-channel Multi-party Meeting

This paper describes a spatial-aware speaker diarization system for the multi-channel multi-party meeting. The diarization system obtains direction information of speaker by microphone array. Speaker spatial embedding is generated by xvector and s-vector derived from superdirective beamforming (SDB) which makes the embedding more robust. Specifically, we propose a novel multi-channel sequence-to-sequence neural network architecture named discriminative multi-stream neural network (DMSNet) which consists of attention superdirective beamforming (ASDB) block and Conformer encoder. The proposed ASDB is a self-adapted channel-wise block that extracts the latent spatial features of array audios by modeling interdependencies between channels. We explore DMSNet to address overlapped speech problem on multi-channel audio and achieve 93.53% accuracy on evaluation set. By performing DMSNet based overlapped speech detection (OSD) module, the diarization error rate (DER) of cluster-based diarization system decrease significantly from 13.45% to 7.64%.

Deep Representation Decomposition for Rate-Invariant Speaker Verification

While promising performance for speaker verification has been achieved by deep speaker embeddings, the advantage would reduce in the case of speaking-style variability. Speaking rate mismatch is often observed in practical speaker verification systems, which may actually degrade the system performance. To reduce intra-class discrepancy caused by speaking rate, we propose a deep representation decomposition approach with adversarial learning to learn speaking rate-invariant speaker embeddings. Specifically, adopting an attention block, we decompose the original embedding into an identity-related component and a rate-related component through multi-task training. Additionally, to reduce the latent relationship between the two decomposed components, we further propose a cosine mapping block to train the parameters adversarially to minimize the cosine similarity between the two decomposed components. As a result, identity-related features become robust to speaking rate and then are used for verification. Experiments are conducted on VoxCeleb1 data and HI-MIA data to demonstrate the effectiveness of our proposed approach.

Graph Convolutional Network Based Semi-Supervised Learning on Multi-Speaker Meeting Data

Unsupervised clustering on speakers is becoming increasingly important for its potential uses in semi-supervised learning. In reality, we are often presented with enormous amounts of unlabeled data from multi-party meetings and discussions. An effective unsupervised clustering approach would allow us to significantly increase the amount of training data without additional costs for annotations. Recently, methods based on graph convolutional networks (GCN) have received growing attention for unsupervised clustering, as these methods exploit the connectivity patterns between nodes to improve learning performance. In this work, we present a GCN-based approach for semi-supervised learning. Given a pre-trained embedding extractor, a graph convolutional network is trained on the labeled data and clusters unlabeled data with "pseudo-labels". We present a self-correcting training mechanism that iteratively runs the cluster-train-correct process on pseudo-labels. We show that this proposed approach effectively uses unlabeled data and improves speaker recognition accuracy.