Multi-Level Graph Contrastive Learning

Graph representation learning has attracted a surge of interest recently, whose target at learning discriminant embedding for each node in the graph. Most of these representation methods focus on supervised learning and heavily depend on label information. However, annotating graphs are expensive to obtain in the real world, especially in specialized domains (i.e. biology), as it needs the annotator to have the domain knowledge to label the graph. To approach this problem, self-supervised learning provides a feasible solution for graph representation learning. In this paper, we propose a Multi-Level Graph Contrastive Learning (MLGCL) framework for learning robust representation of graph data by contrasting space views of graphs. Specifically, we introduce a novel contrastive view - topological and feature space views. The original graph is first-order approximation structure and contains uncertainty or error, while the $k$NN graph generated by encoding features preserves high-order proximity. Thus $k$NN graph generated by encoding features not only provide a complementary view, but is more suitable to GNN encoder to extract discriminant representation. Furthermore, we develop a multi-level contrastive mode to preserve the local similarity and semantic similarity of graph-structured data simultaneously. Extensive experiments indicate MLGCL achieves promising results compared with the existing state-of-the-art graph representation learning methods on seven datasets.

Continual Learning for Fake Audio Detection

Fake audio attack becomes a major threat to the speaker verification system. Although current detection approaches have achieved promising results on dataset-specific scenarios, they encounter difficulties on unseen spoofing data. Fine-tuning and retraining from scratch have been applied to incorporate new data. However, fine-tuning leads to performance degradation on previous data. Retraining takes a lot of time and computation resources. Besides, previous data are unavailable due to privacy in some situations. To solve the above problems, this paper proposes detecting fake without forgetting, a continual-learning-based method, to make the model learn new spoofing attacks incrementally. A knowledge distillation loss is introduced to loss function to preserve the memory of original model. Supposing the distribution of genuine voice is consistent among different scenarios, an extra embedding similarity loss is used as another constraint to further do a positive sample alignment. Experiments are conducted on the ASVspoof2019 dataset. The results show that our proposed method outperforms fine-tuning by the relative reduction of average equal error rate up to 81.62%.

Half-Truth: A Partially Fake Audio Detection Dataset

Diverse promising datasets have been designed to hold back the development of fake audio detection, such as ASVspoof databases. However, previous datasets ignore an attacking situation, in which the hacker hides some small fake clips in real speech audio. This poses a serious threat since that it is difficult to distinguish the small fake clip from the whole speech utterance. Therefore, this paper develops such a dataset for half-truth audio detection (HAD). Partially fake audio in the HAD dataset involves only changing a few words in an utterance.The audio of the words is generated with the very latest state-of-the-art speech synthesis technology. We can not only detect fake uttrances but also localize manipulated regions in a speech using this dataset. Some benchmark results are presented on this dataset. The results show that partially fake audio presents much more challenging than fully fake audio for fake audio detection.

FSR: Accelerating the Inference Process of Transducer-Based Models by Applying Fast-Skip Regularization

Transducer-based models, such as RNN-Transducer and transformer-transducer, have achieved great success in speech recognition. A typical transducer model decodes the output sequence conditioned on the current acoustic state and previously predicted tokens step by step. Statistically, The number of blank tokens in the prediction results accounts for nearly 90\% of all tokens. It takes a lot of computation and time to predict the blank tokens, but only the non-blank tokens will appear in the final output sequence. Therefore, we propose a method named fast-skip regularization, which tries to align the blank position predicted by a transducer with that predicted by a CTC model. During the inference, the transducer model can predict the blank tokens in advance by a simple CTC project layer without many complicated forward calculations of the transducer decoder and then skip them, which will reduce the computation and improve the inference speed greatly. All experiments are conducted on a public Chinese mandarin dataset AISHELL-1. The results show that the fast-skip regularization can indeed help the transducer model learn the blank position alignments. Besides, the inference with fast-skip can be speeded up nearly 4 times with only a little performance degradation.

TSNAT: Two-Step Non-Autoregressvie Transformer Models for Speech Recognition

The autoregressive (AR) models, such as attention-based encoder-decoder models and RNN-Transducer, have achieved great success in speech recognition. They predict the output sequence conditioned on the previous tokens and acoustic encoded states, which is inefficient on GPUs. The non-autoregressive (NAR) models can get rid of the temporal dependency between the output tokens and predict the entire output tokens in at least one step. However, the NAR model still faces two major problems. On the one hand, there is still a great gap in performance between the NAR models and the advanced AR models. On the other hand, it's difficult for most of the NAR models to train and converge. To address these two problems, we propose a new model named the two-step non-autoregressive transformer(TSNAT), which improves the performance and accelerating the convergence of the NAR model by learning prior knowledge from a parameters-sharing AR model. Furthermore, we introduce the two-stage method into the inference process, which improves the model performance greatly. All the experiments are conducted on a public Chinese mandarin dataset ASIEHLL-1. The results show that the TSNAT can achieve a competitive performance with the AR model and outperform many complicated NAR models.

Fast End-to-End Speech Recognition via a Non-Autoregressive Model and Cross-Modal Knowledge Transferring from BERT

Attention-based encoder-decoder (AED) models have achieved promising performance in speech recognition. However, because the decoder predicts text tokens (such as characters or words) in an autoregressive manner, it is difficult for an AED model to predict all tokens in parallel. This makes the inference speed relatively slow. We believe that because the encoder already captures the whole speech utterance, which has the token-level relationship implicitly, we can predict a token without explicitly autoregressive language modeling. When the prediction of a token does not rely on other tokens, the parallel prediction of all tokens in the sequence is realizable. Based on this idea, we propose a non-autoregressive speech recognition model called LASO (Listen Attentively, and Spell Once). The model consists of an encoder, a decoder, and a position dependent summarizer (PDS). The three modules are based on basic attention blocks. The encoder extracts high-level representations from the speech. The PDS uses positional encodings corresponding to tokens to convert the acoustic representations into token-level representations. The decoder further captures token-level relationships with the self-attention mechanism. At last, the probability distribution on the vocabulary is computed for each token position. Therefore, speech recognition is re-formulated as a position-wise classification problem. Further, we propose a cross-modal transfer learning method to refine semantics from a large-scale pre-trained language model BERT for improving the performance.

Fast End-to-End Speech Recognition via Non-Autoregressive Models and Cross-Modal Knowledge Transferring from BERT

Attention-based encoder-decoder (AED) models have achieved promising performance in speech recognition. However, because the decoder predicts text tokens (such as characters or words) in an autoregressive manner, it is difficult for an AED model to predict all tokens in parallel. This makes the inference speed relatively slow. We believe that because the encoder already captures the whole speech utterance, which has the token-level relationship implicitly, we can predict a token without explicitly autoregressive language modeling. When the prediction of a token does not rely on other tokens, the parallel prediction of all tokens in the sequence is realizable. Based on this idea, we propose a non-autoregressive speech recognition model called LASO (Listen Attentively, and Spell Once). The model consists of an encoder, a decoder, and a position dependent summarizer (PDS). The three modules are based on basic attention blocks. The encoder extracts high-level representations from the speech. The PDS uses positional encodings corresponding to tokens to convert the acoustic representations into token-level representations. The decoder further captures token-level relationships with the self-attention mechanism. At last, the probability distribution on the vocabulary is computed for each token position. Therefore, speech recognition is re-formulated as a position-wise classification problem. Further, we propose a cross-modal transfer learning method to refine semantics from a large-scale pre-trained language model BERT for improving the performance.

Tucker decomposition-based Temporal Knowledge Graph Completion

Knowledge graphs have been demonstrated to be an effective tool for numerous intelligent applications. However, a large amount of valuable knowledge still exists implicitly in the knowledge graphs. To enrich the existing knowledge graphs, recent years witness that many algorithms for link prediction and knowledge graphs embedding have been designed to infer new facts. But most of these studies focus on the static knowledge graphs and ignore the temporal information that reflects the validity of knowledge. Developing the model for temporal knowledge graphs completion is an increasingly important task. In this paper, we build a new tensor decomposition model for temporal knowledge graphs completion inspired by the Tucker decomposition of order 4 tensor. We demonstrate that the proposed model is fully expressive and report state-of-the-art results for several public benchmarks. Additionally, we present several regularization schemes to improve the strategy and study their impact on the proposed model. Experimental studies on three temporal datasets (i.e. ICEWS2014, ICEWS2005-15, GDELT) justify our design and demonstrate that our model outperforms baselines with an explicit margin on link prediction task.

Self-supervised Graph Representation Learning via Bootstrapping

Graph neural networks~(GNNs) apply deep learning techniques to graph-structured data and have achieved promising performance in graph representation learning. However, existing GNNs rely heavily on enough labels or well-designed negative samples. To address these issues, we propose a new self-supervised graph representation method: deep graph bootstrapping~(DGB). DGB consists of two neural networks: online and target networks, and the input of them are different augmented views of the initial graph. The online network is trained to predict the target network while the target network is updated with a slow-moving average of the online network, which means the online and target networks can learn from each other. As a result, the proposed DGB can learn graph representation without negative examples in an unsupervised manner. In addition, we summarize three kinds of augmentation methods for graph-structured data and apply them to the DGB. Experiments on the benchmark datasets show the DGB performs better than the current state-of-the-art methods and how the augmentation methods affect the performances.