Liang Chen

Chen Liang, Simiao Zuo, Qingru Zhang et al. · gatech, microsoft-research

Layer-wise distillation is a powerful tool to compress large models (i.e. teacher models) into small ones (i.e., student models). The student distills knowledge from the teacher by mimicking the hidden representations of the teacher at every intermediate layer. However, layer-wise distillation is difficult. Since the student has a smaller model capacity than the teacher, it is often under-fitted. Furthermore, the hidden representations of the teacher contain redundant information that the student does not necessarily need for the target task's learning. To address these challenges, we propose a novel Task-aware layEr-wise Distillation (TED). TED designs task-aware filters to align the hidden representations of the student and the teacher at each layer. The filters select the knowledge that is useful for the target task from the hidden representations. As such, TED reduces the knowledge gap between the two models and helps the student to fit better on the target task. We evaluate TED in two scenarios: continual pre-training and fine-tuning. TED demonstrates significant and consistent improvements over existing distillation methods in both scenarios. Code is available at https://github.com/cliang1453/task-aware-distillation.

Qingru Zhang, Simiao Zuo, Chen Liang et al. · gatech, microsoft-research

Large Transformer-based models have exhibited superior performance in various natural language processing and computer vision tasks. However, these models contain enormous amounts of parameters, which restrict their deployment to real-world applications. To reduce the model size, researchers prune these models based on the weights' importance scores. However, such scores are usually estimated on mini-batches during training, which incurs large variability/uncertainty due to mini-batch sampling and complicated training dynamics. As a result, some crucial weights could be pruned by commonly used pruning methods because of such uncertainty, which makes training unstable and hurts generalization. To resolve this issue, we propose PLATON, which captures the uncertainty of importance scores by upper confidence bound (UCB) of importance estimation. In particular, for the weights with low importance scores but high uncertainty, PLATON tends to retain them and explores their capacity. We conduct extensive experiments with several Transformer-based models on natural language understanding, question answering and image classification to validate the effectiveness of PLATON. Results demonstrate that PLATON manifests notable improvement under different sparsity levels. Our code is publicly available at https://github.com/QingruZhang/PLATON.

Simiao Zuo, Qingru Zhang, Chen Liang et al. · gatech, microsoft-research

Pre-trained language models have demonstrated superior performance in various natural language processing tasks. However, these models usually contain hundreds of millions of parameters, which limits their practicality because of latency requirements in real-world applications. Existing methods train small compressed models via knowledge distillation. However, performance of these small models drops significantly compared with the pre-trained models due to their reduced model capacity. We propose MoEBERT, which uses a Mixture-of-Experts structure to increase model capacity and inference speed. We initialize MoEBERT by adapting the feed-forward neural networks in a pre-trained model into multiple experts. As such, representation power of the pre-trained model is largely retained. During inference, only one of the experts is activated, such that speed can be improved. We also propose a layer-wise distillation method to train MoEBERT. We validate the efficiency and effectiveness of MoEBERT on natural language understanding and question answering tasks. Results show that the proposed method outperforms existing task-specific distillation algorithms. For example, our method outperforms previous approaches by over 2% on the MNLI (mismatched) dataset. Our code is publicly available at https://github.com/SimiaoZuo/MoEBERT.

Liang Chen, Runxin Xu, Baobao Chang · pku

Label smoothing and vocabulary sharing are two widely used techniques in neural machine translation models. However, we argue that simply applying both techniques can be conflicting and even leads to sub-optimal performance. When allocating smoothed probability, original label smoothing treats the source-side words that would never appear in the target language equally to the real target-side words, which could bias the translation model. To address this issue, we propose Masked Label Smoothing (MLS), a new mechanism that masks the soft label probability of source-side words to zero. Simple yet effective, MLS manages to better integrate label smoothing with vocabulary sharing. Our extensive experiments show that MLS consistently yields improvement over original label smoothing on different datasets, including bilingual and multilingual translation from both translation quality and model's calibration. Our code is released at https://github.com/PKUnlp-icler/MLS

Jintang Li, Ruofan Wu, Wangbin Sun et al.

The last years have witnessed the emergence of a promising self-supervised learning strategy, referred to as masked autoencoding. However, there is a lack of theoretical understanding of how masking matters on graph autoencoders (GAEs). In this work, we present masked graph autoencoder (MaskGAE), a self-supervised learning framework for graph-structured data. Different from standard GAEs, MaskGAE adopts masked graph modeling (MGM) as a principled pretext task - masking a portion of edges and attempting to reconstruct the missing part with partially visible, unmasked graph structure. To understand whether MGM can help GAEs learn better representations, we provide both theoretical and empirical evidence to comprehensively justify the benefits of this pretext task. Theoretically, we establish close connections between GAEs and contrastive learning, showing that MGM significantly improves the self-supervised learning scheme of GAEs. Empirically, we conduct extensive experiments on a variety of graph benchmarks, demonstrating the superiority of MaskGAE over several state-of-the-arts on both link prediction and node classification tasks.

Liang Chen, Bofei Gao, Baobao Chang · pku

In this paper, we provide a detailed description of our system at CAMRP-2022 evaluation. We firstly propose a two-stage method to conduct Chinese AMR Parsing with alignment generation, which includes Concept-Prediction and Relation-Prediction stages. Our model achieves 0.7756 and 0.7074 Align-Smatch F1 scores on the CAMR 2.0 test set and the blind-test set of CAMRP-2022 individually. We also analyze the result and the limitation such as the error propagation and class imbalance problem we conclude in the current method. Code and the trained models are released at https://github.com/PKUnlp-icler/Two-Stage-CAMRP for reproduction.

Chen Liang, Wenguan Wang, Jiaxu Miao et al.

Recent advances in semi-supervised semantic segmentation have been heavily reliant on pseudo labeling to compensate for limited labeled data, disregarding the valuable relational knowledge among semantic concepts. To bridge this gap, we devise LogicDiag, a brand new neural-logic semi-supervised learning framework. Our key insight is that conflicts within pseudo labels, identified through symbolic knowledge, can serve as strong yet commonly ignored learning signals. LogicDiag resolves such conflicts via reasoning with logic-induced diagnoses, enabling the recovery of (potentially) erroneous pseudo labels, ultimately alleviating the notorious error accumulation problem. We showcase the practical application of LogicDiag in the data-hungry segmentation scenario, where we formalize the structured abstraction of semantic concepts as a set of logic rules. Extensive experiments on three standard semi-supervised semantic segmentation benchmarks demonstrate the effectiveness and generality of LogicDiag. Moreover, LogicDiag highlights the promising opportunities arising from the systematic integration of symbolic reasoning into the prevalent statistical, neural learning approaches.

Jintang Li, Jiaying Peng, Liang Chen et al.

Recent studies demonstrate that Graph Neural Networks (GNNs) are vulnerable to slight but adversarially designed perturbations, known as adversarial examples. To address this issue, robust training methods against adversarial examples have received considerable attention in the literature. \emph{Adversarial Training (AT)} is a successful approach to learning a robust model using adversarially perturbed training samples. Existing AT methods on GNNs typically construct adversarial perturbations in terms of graph structures or node features. However, they are less effective and fraught with challenges on graph data due to the discreteness of graph structure and the relationships between connected examples. In this work, we seek to address these challenges and propose Spectral Adversarial Training (SAT), a simple yet effective adversarial training approach for GNNs. SAT first adopts a low-rank approximation of the graph structure based on spectral decomposition, and then constructs adversarial perturbations in the spectral domain rather than directly manipulating the original graph structure. To investigate its effectiveness, we employ SAT on three widely used GNNs. Experimental results on four public graph datasets demonstrate that SAT significantly improves the robustness of GNNs against adversarial attacks without sacrificing classification accuracy and training efficiency.

Bingzhe Wu, Jintang Li, Junchi Yu et al.

Deep graph learning has achieved remarkable progresses in both business and scientific areas ranging from finance and e-commerce, to drug and advanced material discovery. Despite these progresses, how to ensure various deep graph learning algorithms behave in a socially responsible manner and meet regulatory compliance requirements becomes an emerging problem, especially in risk-sensitive domains. Trustworthy graph learning (TwGL) aims to solve the above problems from a technical viewpoint. In contrast to conventional graph learning research which mainly cares about model performance, TwGL considers various reliability and safety aspects of the graph learning framework including but not limited to robustness, explainability, and privacy. In this survey, we provide a comprehensive review of recent leading approaches in the TwGL field from three dimensions, namely, reliability, explainability, and privacy protection. We give a general categorization for existing work and review typical work for each category. To give further insights for TwGL research, we provide a unified view to inspect previous works and build the connection between them. We also point out some important open problems remaining to be solved in the future developments of TwGL.

Jintang Li, Jie Liao, Ruofan Wu et al.

Graph convolutional networks (GCNs) have been shown to be vulnerable to small adversarial perturbations, which becomes a severe threat and largely limits their applications in security-critical scenarios. To mitigate such a threat, considerable research efforts have been devoted to increasing the robustness of GCNs against adversarial attacks. However, current defense approaches are typically designed to prevent GCNs from untargeted adversarial attacks and focus on overall performance, making it challenging to protect important local nodes from more powerful targeted adversarial attacks. Additionally, a trade-off between robustness and performance is often made in existing research. Such limitations highlight the need for developing an effective and efficient approach that can defend local nodes against targeted attacks, without compromising the overall performance of GCNs. In this work, we present a simple yet effective method, named Graph Universal Adversarial Defense (GUARD). Unlike previous works, GUARD protects each individual node from attacks with a universal defensive patch, which is generated once and can be applied to any node (node-agnostic) in a graph. GUARD is fast, straightforward to implement without any change to network architecture nor any additional parameters, and is broadly applicable to any GCNs. Extensive experiments on four benchmark datasets demonstrate that GUARD significantly improves robustness for several established GCNs against multiple adversarial attacks and outperforms state-of-the-art defense methods by large margins.

Yuansheng Wang, Wangbin Sun, Kun Xu et al.

Graph contrastive learning (GCL), as a popular approach to graph self-supervised learning, has recently achieved a non-negligible effect. To achieve superior performance, the majority of existing GCL methods elaborate on graph data augmentation to construct appropriate contrastive pairs. However, existing methods place more emphasis on the complex graph data augmentation which requires extra time overhead, and pay less attention to developing contrastive schemes specific to encoder characteristics. We argue that a better contrastive scheme should be tailored to the characteristics of graph neural networks (e.g., neighborhood aggregation) and propose a simple yet effective method named FastGCL. Specifically, by constructing weighted-aggregated and non-aggregated neighborhood information as positive and negative samples respectively, FastGCL identifies the potential semantic information of data without disturbing the graph topology and node attributes, resulting in faster training and convergence speeds. Extensive experiments have been conducted on node classification and graph classification tasks, showing that FastGCL has competitive classification performance and significant training speedup compared to existing state-of-the-art methods.

Jintang Li, Tao Xie, Liang Chen et al.

Recent studies have shown that graph neural networks (GNNs) are vulnerable against perturbations due to lack of robustness and can therefore be easily fooled. Currently, most works on attacking GNNs are mainly using gradient information to guide the attack and achieve outstanding performance. However, the high complexity of time and space makes them unmanageable for large scale graphs and becomes the major bottleneck that prevents the practical usage. We argue that the main reason is that they have to use the whole graph for attacks, resulting in the increasing time and space complexity as the data scale grows. In this work, we propose an efficient Simplified Gradient-based Attack (SGA) method to bridge this gap. SGA can cause the GNNs to misclassify specific target nodes through a multi-stage attack framework, which needs only a much smaller subgraph. In addition, we present a practical metric named Degree Assortativity Change (DAC) to measure the impacts of adversarial attacks on graph data. We evaluate our attack method on four real-world graph networks by attacking several commonly used GNNs. The experimental results demonstrate that SGA can achieve significant time and memory efficiency improvements while maintaining competitive attack performance compared to state-of-art attack techniques. Codes are available via: https://github.com/EdisonLeeeee/SGAttack.

Liang Chen, Yanchun Liang, Xiaohu Shi et al.

Time Delay Neural Network (TDNN) is a well-performing structure for DNN-based speaker recognition systems. In this paper we introduce a novel structure Crossed-Time Delay Neural Network (CTDNN) to enhance the performance of current TDNN. Inspired by the multi-filters setting of convolution layer from convolution neural network, we set multiple time delay units each with different context size at the bottom layer and construct a multilayer parallel network. The proposed CTDNN gives significant improvements over original TDNN on both speaker verification and identification tasks. It outperforms in VoxCeleb1 dataset in verification experiment with a 2.6% absolute Equal Error Rate improvement. In few shots condition CTDNN reaches 90.4% identification accuracy, which doubles the identification accuracy of original TDNN. We also compare the proposed CTDNN with another new variant of TDNN, FTDNN, which shows that our model has a 36% absolute identification accuracy improvement under few shots condition and can better handle training of a larger batch in a shorter training time, which better utilize the calculation resources. The code of the new model is released at https://github.com/chenllliang/CTDNN

Jintang Li, Bingzhe Wu, Chengbin Hou et al.

Deep graph learning (DGL) has achieved remarkable progress in both business and scientific areas ranging from finance and e-commerce to drug and advanced material discovery. Despite the progress, applying DGL to real-world applications faces a series of reliability threats including inherent noise, distribution shift, and adversarial attacks. This survey aims to provide a comprehensive review of recent advances for improving the reliability of DGL algorithms against the above threats. In contrast to prior related surveys which mainly focus on adversarial attacks and defense, our survey covers more reliability-related aspects of DGL, i.e., inherent noise and distribution shift. Additionally, we discuss the relationships among above aspects and highlight some important issues to be explored in future research.

Chen Liang, Haoming Jiang, Simiao Zuo et al.

Recent research has shown the existence of significant redundancy in large Transformer models. One can prune the redundant parameters without significantly sacrificing the generalization performance. However, we question whether the redundant parameters could have contributed more if they were properly trained. To answer this question, we propose a novel training strategy that encourages all parameters to be trained sufficiently. Specifically, we adaptively adjust the learning rate for each parameter according to its sensitivity, a robust gradient-based measure reflecting this parameter's contribution to the model performance. A parameter with low sensitivity is redundant, and we improve its fitting by increasing its learning rate. In contrast, a parameter with high sensitivity is well-trained, and we regularize it by decreasing its learning rate to prevent further overfitting. We conduct extensive experiments on natural language understanding, neural machine translation, and image classification to demonstrate the effectiveness of the proposed schedule. Analysis shows that the proposed schedule indeed reduces the redundancy and improves generalization performance.

Chenlin Shen, Guangda Huzhang, Yuhang Zhou et al.

To approach different business objectives, online traffic shaping algorithms aim at improving exposures of a target set of items, such as boosting the growth of new commodities. Generally, these algorithms assume that the utility of each user-item pair can be accessed via a well-trained conversion rate prediction model. However, for real E-Commerce platforms, there are unavoidable factors preventing us from learning such an accurate model. In order to break the heavy dependence on accurate inputs of the utility, we propose a general online traffic shaping protocol for online E-Commerce applications. In our framework, we approximate the function mapping the bonus scores, which generally are the only method to influence the ranking result in the traffic shaping problem, to the numbers of exposures and purchases. Concretely, we approximate the above function by a class of the piece-wise linear function constructed on the convex hull of the explored data points. Moreover, we reformulate the online traffic shaping problem as linear programming where these piece-wise linear functions are embedded into both the objective and constraints. Our algorithm can straightforwardly optimize the linear programming in the prime space, and its solution can be simply applied by a stochastic strategy to fulfill the optimized objective and the constraints in expectation. Finally, the online A/B test shows our proposed algorithm steadily outperforms the previous industrial level traffic shaping algorithm.

Simiao Zuo, Yue Yu, Chen Liang et al.

Self-training achieves enormous success in various semi-supervised and weakly-supervised learning tasks. The method can be interpreted as a teacher-student framework, where the teacher generates pseudo-labels, and the student makes predictions. The two models are updated alternatingly. However, such a straightforward alternating update rule leads to training instability. This is because a small change in the teacher may result in a significant change in the student. To address this issue, we propose DRIFT, short for differentiable self-training, that treats teacher-student as a Stackelberg game. In this game, a leader is always in a more advantageous position than a follower. In self-training, the student contributes to the prediction performance, and the teacher controls the training process by generating pseudo-labels. Therefore, we treat the student as the leader and the teacher as the follower. The leader procures its advantage by acknowledging the follower's strategy, which involves differentiable pseudo-labels and differentiable sample weights. Consequently, the leader-follower interaction can be effectively captured via Stackelberg gradient, obtained by differentiating the follower's strategy. Experimental results on semi- and weakly-supervised classification and named entity recognition tasks show that our model outperforms existing approaches by large margins.

Liang Chen, Jintang Li, Qibiao Peng et al.

Recent studies have shown that Graph Convolutional Networks (GCNs) are vulnerable to adversarial attacks on the graph structure. Although multiple works have been proposed to improve their robustness against such structural adversarial attacks, the reasons for the success of the attacks remain unclear. In this work, we theoretically and empirically demonstrate that structural adversarial examples can be attributed to the non-robust aggregation scheme (i.e., the weighted mean) of GCNs. Specifically, our analysis takes advantage of the breakdown point which can quantitatively measure the robustness of aggregation schemes. The key insight is that weighted mean, as the basic design of GCNs, has a low breakdown point and its output can be dramatically changed by injecting a single edge. We show that adopting the aggregation scheme with a high breakdown point (e.g., median or trimmed mean) could significantly enhance the robustness of GCNs against structural attacks. Extensive experiments on four real-world datasets demonstrate that such a simple but effective method achieves the best robustness performance compared to state-of-the-art models.

Carl Yang, Haonan Wang, Ke Zhang et al.

Many data mining and analytical tasks rely on the abstraction of networks (graphs) to summarize relational structures among individuals (nodes). Since relational data are often sensitive, we aim to seek effective approaches to generate utility-preserved yet privacy-protected structured data. In this paper, we leverage the differential privacy (DP) framework to formulate and enforce rigorous privacy constraints on deep graph generation models, with a focus on edge-DP to guarantee individual link privacy. In particular, we enforce edge-DP by injecting proper noise to the gradients of a link reconstruction-based graph generation model, while ensuring data utility by improving structure learning with structure-oriented graph discrimination. Extensive experiments on two real-world network datasets show that our proposed DPGGAN model is able to generate graphs with effectively preserved global structure and rigorously protected individual link privacy.