Learning to Behave Like Clean Speech: Dual-Branch Knowledge Distillation for Noise-Robust Fake Audio Detection

Most research in fake audio detection (FAD) focuses on improving performance on standard noise-free datasets. However, in actual situations, there is usually noise interference, which will cause significant performance degradation in FAD systems. To improve the noise robustness, we propose a dual-branch knowledge distillation fake audio detection (DKDFAD) method. Specifically, a parallel data flow of the clean teacher branch and the noisy student branch is designed, and interactive fusion and response-based teacher-student paradigms are proposed to guide the training of noisy data from the data distribution and decision-making perspectives. In the noise branch, speech enhancement is first introduced for denoising, which reduces the interference of strong noise. The proposed interactive fusion combines denoising features and noise features to reduce the impact of speech distortion and seek consistency with the data distribution of clean branch. The teacher-student paradigm maps the student's decision space to the teacher's decision space, making noisy speech behave as clean. In addition, a joint training method is used to optimize the two branches to achieve global optimality. Experimental results based on multiple datasets show that the proposed method performs well in noisy environments and maintains performance in cross-dataset experiments.

Learning Speech Representation From Contrastive Token-Acoustic Pretraining

For fine-grained generation and recognition tasks such as minimally-supervised text-to-speech (TTS), voice conversion (VC), and automatic speech recognition (ASR), the intermediate representations extracted from speech should serve as a "bridge" between text and acoustic information, containing information from both modalities. The semantic content is emphasized, while the paralinguistic information such as speaker identity and acoustic details should be de-emphasized. However, existing methods for extracting fine-grained intermediate representations from speech suffer from issues of excessive redundancy and dimension explosion. Contrastive learning is a good method for modeling intermediate representations from two modalities. However, existing contrastive learning methods in the audio field focus on extracting global descriptive information for downstream audio classification tasks, making them unsuitable for TTS, VC, and ASR tasks. To address these issues, we propose a method named "Contrastive Token-Acoustic Pretraining (CTAP)", which uses two encoders to bring phoneme and speech into a joint multimodal space, learning how to connect phoneme and speech at the frame level. The CTAP model is trained on 210k speech and phoneme text pairs, achieving minimally-supervised TTS, VC, and ASR. The proposed CTAP method offers a promising solution for fine-grained generation and recognition downstream tasks in speech processing.

CPSP: Learning Speech Concepts From Phoneme Supervision

For fine-grained generation and recognition tasks such as minimally-supervised text-to-speech (TTS), voice conversion (VC), and automatic speech recognition (ASR), the intermediate representation extracted from speech should contain information that is between text coding and acoustic coding. The linguistic content is salient, while the paralinguistic information such as speaker identity and acoustic details should be removed. However, existing methods for extracting fine-grained intermediate representations from speech suffer from issues of excessive redundancy and dimension explosion. Additionally, existing contrastive learning methods in the audio field focus on extracting global descriptive information for downstream audio classification tasks, making them unsuitable for TTS, VC, and ASR tasks. To address these issues, we propose a method named Contrastive Phoneme-Speech Pretraining (CPSP), which uses three encoders, one decoder, and contrastive learning to bring phoneme and speech into a joint multimodal space, learning how to connect phoneme and speech at the frame level. The CPSP model is trained on 210k speech and phoneme text pairs, achieving minimally-supervised TTS, VC, and ASR. The proposed CPSP method offers a promising solution for fine-grained generation and recognition downstream tasks in speech processing. We provide a website with audio samples.

Minimally-Supervised Speech Synthesis with Conditional Diffusion Model and Language Model: A Comparative Study of Semantic Coding

Recently, there has been a growing interest in text-to-speech (TTS) methods that can be trained with minimal supervision by combining two types of discrete speech representations and using two sequence-to-sequence tasks to decouple TTS. To address the challenges associated with high dimensionality and waveform distortion in discrete representations, we propose Diff-LM-Speech, which models semantic embeddings into mel-spectrogram based on diffusion models and introduces a prompt encoder structure based on variational autoencoders and prosody bottlenecks to improve prompt representation capabilities. Autoregressive language models often suffer from missing and repeated words, while non-autoregressive frameworks face expression averaging problems due to duration prediction models. To address these issues, we propose Tetra-Diff-Speech, which designs a duration diffusion model to achieve diverse prosodic expressions. While we expect the information content of semantic coding to be between that of text and acoustic coding, existing models extract semantic coding with a lot of redundant information and dimensionality explosion. To verify that semantic coding is not necessary, we propose Tri-Diff-Speech. Experimental results show that our proposed methods outperform baseline methods. We provide a website with audio samples.

Low-rank Adaptation Method for Wav2vec2-based Fake Audio Detection

Self-supervised speech models are a rapidly developing research topic in fake audio detection. Many pre-trained models can serve as feature extractors, learning richer and higher-level speech features. However,when fine-tuning pre-trained models, there is often a challenge of excessively long training times and high memory consumption, and complete fine-tuning is also very expensive. To alleviate this problem, we apply low-rank adaptation(LoRA) to the wav2vec2 model, freezing the pre-trained model weights and injecting a trainable rank-decomposition matrix into each layer of the transformer architecture, greatly reducing the number of trainable parameters for downstream tasks. Compared with fine-tuning with Adam on the wav2vec2 model containing 317M training parameters, LoRA achieved similar performance by reducing the number of trainable parameters by 198 times.

Adaptive Fake Audio Detection with Low-Rank Model Squeezing

The rapid advancement of spoofing algorithms necessitates the development of robust detection methods capable of accurately identifying emerging fake audio. Traditional approaches, such as finetuning on new datasets containing these novel spoofing algorithms, are computationally intensive and pose a risk of impairing the acquired knowledge of known fake audio types. To address these challenges, this paper proposes an innovative approach that mitigates the limitations associated with finetuning. We introduce the concept of training low-rank adaptation matrices tailored specifically to the newly emerging fake audio types. During the inference stage, these adaptation matrices are combined with the existing model to generate the final prediction output. Extensive experimentation is conducted to evaluate the efficacy of the proposed method. The results demonstrate that our approach effectively preserves the prediction accuracy of the existing model for known fake audio types. Furthermore, our approach offers several advantages, including reduced storage memory requirements and lower equal error rates compared to conventional finetuning methods, particularly on specific spoofing algorithms.

ADD 2023: the Second Audio Deepfake Detection Challenge

Audio deepfake detection is an emerging topic in the artificial intelligence community. The second Audio Deepfake Detection Challenge (ADD 2023) aims to spur researchers around the world to build new innovative technologies that can further accelerate and foster research on detecting and analyzing deepfake speech utterances. Different from previous challenges (e.g. ADD 2022), ADD 2023 focuses on surpassing the constraints of binary real/fake classification, and actually localizing the manipulated intervals in a partially fake speech as well as pinpointing the source responsible for generating any fake audio. Furthermore, ADD 2023 includes more rounds of evaluation for the fake audio game sub-challenge. The ADD 2023 challenge includes three subchallenges: audio fake game (FG), manipulation region location (RL) and deepfake algorithm recognition (AR). This paper describes the datasets, evaluation metrics, and protocols. Some findings are also reported in audio deepfake detection tasks.

TO-Rawnet: Improving RawNet with TCN and Orthogonal Regularization for Fake Audio Detection

Current fake audio detection relies on hand-crafted features, which lose information during extraction. To overcome this, recent studies use direct feature extraction from raw audio signals. For example, RawNet is one of the representative works in end-to-end fake audio detection. However, existing work on RawNet does not optimize the parameters of the Sinc-conv during training, which limited its performance. In this paper, we propose to incorporate orthogonal convolution into RawNet, which reduces the correlation between filters when optimizing the parameters of Sinc-conv, thus improving discriminability. Additionally, we introduce temporal convolutional networks (TCN) to capture long-term dependencies in speech signals. Experiments on the ASVspoof 2019 show that the Our TO-RawNet system can relatively reduce EER by 66.09\% on logical access scenario compared with the RawNet, demonstrating its effectiveness in detecting fake audio attacks.

UnifySpeech: A Unified Framework for Zero-shot Text-to-Speech and Voice Conversion

Text-to-speech (TTS) and voice conversion (VC) are two different tasks both aiming at generating high quality speaking voice according to different input modality. Due to their similarity, this paper proposes UnifySpeech, which brings TTS and VC into a unified framework for the first time. The model is based on the assumption that speech can be decoupled into three independent components: content information, speaker information, prosody information. Both TTS and VC can be regarded as mining these three parts of information from the input and completing the reconstruction of speech. For TTS, the speech content information is derived from the text, while in VC it's derived from the source speech, so all the remaining units are shared except for the speech content extraction module in the two tasks. We applied vector quantization and domain constrain to bridge the gap between the content domains of TTS and VC. Objective and subjective evaluation shows that by combining the two task, TTS obtains better speaker modeling ability while VC gets hold of impressive speech content decoupling capability.

Emotion Selectable End-to-End Text-based Speech Editing

Text-based speech editing allows users to edit speech by intuitively cutting, copying, and pasting text to speed up the process of editing speech. In the previous work, CampNet (context-aware mask prediction network) is proposed to realize text-based speech editing, significantly improving the quality of edited speech. This paper aims at a new task: adding emotional effect to the editing speech during the text-based speech editing to make the generated speech more expressive. To achieve this task, we propose Emo-CampNet (emotion CampNet), which can provide the option of emotional attributes for the generated speech in text-based speech editing and has the one-shot ability to edit unseen speakers' speech. Firstly, we propose an end-to-end emotion-selectable text-based speech editing model. The key idea of the model is to control the emotion of generated speech by introducing additional emotion attributes based on the context-aware mask prediction network. Secondly, to prevent the emotion of the generated speech from being interfered by the emotional components in the original speech, a neutral content generator is proposed to remove the emotion from the original speech, which is optimized by the generative adversarial framework. Thirdly, two data augmentation methods are proposed to enrich the emotional and pronunciation information in the training set, which can enable the model to edit the unseen speaker's speech. The experimental results that 1) Emo-CampNet can effectively control the emotion of the generated speech in the process of text-based speech editing; And can edit unseen speakers' speech. 2) Detailed ablation experiments further prove the effectiveness of emotional selectivity and data augmentation methods. The demo page is available at https://hairuo55.github.io/Emo-CampNet/