Cheng Wu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Cheng Wu has a total of 82 co-authors. Here are the top co-authors by number of shared publications:
- Shiji Song
Tsinghua University
23 shared publications - Keyou You
Tsinghua University
6 shared publications - Gao Huang
Tsinghua University, Beijing, China
4 shared publications - Rui Zhang
Department of Industrial Engineering and Management, Xiamen University of Technology, Xiamen, China
3 shared publications - Wenjie Shi
Department of Automation, Tsinghua University, Beijing, China
3 shared publications - Feng Jin
Tsinghua University
2 shared publications
Latest Publications
- 12 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 40, 3800-3805
Equivalence analysis of different sequence impedance modeling methods
The sequence impedance has a clear physical meaning and can be directly measured by commercial equipment, so it is widely used for stability analysis in the field of electrical engineering. It can be transformed from dq impedance indirectly or calculated based on Harmonic State Space (HSS) theory directly. This paper analyzes the equivalence of the above two methods. Firstly, the dq impedance model of grid-connected converter is established, and the equivalence of impedance in dq domain and sequence domain is proved by mathematical transformation. Then, modified sequence impedance model of grid-connected converter is established considering the frequency coupling phenomenon. On the other hand, the impedance model can also be obtained from HSS modeling with h order harmonics. Subsequently, the relationship between the modified sequence impedance and HSS impedance can be found. The former is a special case when the latter takes h=2. When considering the coupling effect of higher harmonics, HSS impedance model has higher accuracy. Finally, the accuracy and the equivalence of the above-mentioned two methods are verified by frequency sweep.
- 12 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 40, 3800-3805
Equivalence analysis of different sequence impedance modeling methods
The sequence impedance has a clear physical meaning and can be directly measured by commercial equipment, so it is widely used for stability analysis in the field of electrical engineering. It can be transformed from dq impedance indirectly or calculated based on Harmonic State Space (HSS) theory directly. This paper analyzes the equivalence of the above two methods. Firstly, the dq impedance model of grid-connected converter is established, and the equivalence of impedance in dq domain and sequence domain is proved by mathematical transformation. Then, modified sequence impedance model of grid-connected converter is established considering the frequency coupling phenomenon. On the other hand, the impedance model can also be obtained from HSS modeling with h order harmonics. Subsequently, the relationship between the modified sequence impedance and HSS impedance can be found. The former is a special case when the latter takes h=2. When considering the coupling effect of higher harmonics, HSS impedance model has higher accuracy. Finally, the accuracy and the equivalence of the above-mentioned two methods are verified by frequency sweep.
- Open Access
HybridGNN: Learning Hybrid Representation in Multiplex Heterogeneous Networks
Recently, graph neural networks have shown the superiority of modeling the complex topological structures in heterogeneous network-based recommender systems. Due to the diverse interactions among nodes and abundant semantics emerging from diverse types of nodes and edges, there is a bursting research interest in learning expressive node representations in multiplex heterogeneous networks. One of the most important tasks in recommender systems is to predict the potential connection between two nodes under a specific edge type (i.e., relationship). Although existing studies utilize explicit metapaths to aggregate neighbors, practically they only consider intra-relationship metapaths and thus fail to leverage the potential uplift by inter-relationship information. Moreover, it is not always straightforward to exploit inter-relationship metapaths comprehensively under diverse relationships, especially with the increasing number of node and edge types. In addition, contributions of different relationships between two nodes are difficult to measure. To address the challenges, we propose HybridGNN, an end-to-end GNN model with hybrid aggregation flows and hierarchical attentions to fully utilize the heterogeneity in the multiplex scenarios. Specifically, HybridGNN applies a randomized inter-relationship exploration module to exploit the multiplexity property among different relationships. Then, our model leverages hybrid aggregation flows under intra-relationship metapaths and randomized exploration to learn the rich semantics. To explore the importance of different aggregation flow and take advantage of the multiplexity property, we bring forward a novel hierarchical attention module which leverages both metapath-level attention and relationship-level attention. Extensive experimental results suggest that HybridGNN achieves the best performance compared to several state-of-the-art baselines.
Open AccessHybridGNN: Learning Hybrid Representation in Multiplex Heterogeneous Networks
Recently, graph neural networks have shown the superiority of modeling the complex topological structures in heterogeneous network-based recommender systems. Due to the diverse interactions among nodes and abundant semantics emerging from diverse types of nodes and edges, there is a bursting research interest in learning expressive node representations in multiplex heterogeneous networks. One of the most important tasks in recommender systems is to predict the potential connection between two nodes under a specific edge type (i.e., relationship). Although existing studies utilize explicit metapaths to aggregate neighbors, practically they only consider intra-relationship metapaths and thus fail to leverage the potential uplift by inter-relationship information. Moreover, it is not always straightforward to exploit inter-relationship metapaths comprehensively under diverse relationships, especially with the increasing number of node and edge types. In addition, contributions of different relationships between two nodes are difficult to measure. To address the challenges, we propose HybridGNN, an end-to-end GNN model with hybrid aggregation flows and hierarchical attentions to fully utilize the heterogeneity in the multiplex scenarios. Specifically, HybridGNN applies a randomized inter-relationship exploration module to exploit the multiplexity property among different relationships. Then, our model leverages hybrid aggregation flows under intra-relationship metapaths and randomized exploration to learn the rich semantics. To explore the importance of different aggregation flow and take advantage of the multiplexity property, we bring forward a novel hierarchical attention module which leverages both metapath-level attention and relationship-level attention. Extensive experimental results suggest that HybridGNN achieves the best performance compared to several state-of-the-art baselines.
- Open Access
- 21 June 2022
- Published by Frontiers Media SA in Frontiers in Public Health
- Vol. 10, 903183
Fertility Intention Among Chinese Reproductive Couples During the COVID-19 Outbreak: A Cross-Sectional Study
A decline in the fertility rate has been observed worldwide, which hampers social development severely. Given the impacts of COVID-19 on individuals and society, it is of great significance to investigate the fertility intention of reproductive couples under COVID-19. The convenience sampling method was used to obtain our study sample. The self-administered questionnaire included the following components: sociodemographic characteristics (age, residence, education, occupation, characteristics of the couples, and annual household income), reproductive history (parity, number of children, child gender, and duration of preparing pregnancy), and attitudes toward COVID-19, was distributed online via an applet of WeChat. The results showed that among 4,133 valid questionnaires, 1,091 had fertility intention before COVID-19, whereas 3,042 did not, indicating a fertility intention rate of 26.4% among participating couples. Of the 1,091 couples who had fertility intention before COVID-19, 520 (47.7%) were affected by the outbreak, whereas 571 (52.3%) were not. By multivariable logistic regression analysis, we further found that couples living in Hubei Province, the epicenter in China (OR 2.20, 95% CI 1.35–3.60), and couples who prepared for pregnancy longer before COVID-19 (OR 1.19, 95% CI 1.06–1.33) were more likely to change their fertility intention under the pandemic. In addition, most of the participants reported their fertility intention was affected by the inconvenience of seeking medical service under COVID-19. Therefore, more forms of medical services to provide convenience for patients might be effective ways to reverse the declined fertility intention rate in facing COVID-19.
- 21 June 2022
- Published by Frontiers Media SA in Frontiers in Public Health
- Vol. 10, 903183
Open AccessFertility Intention Among Chinese Reproductive Couples During the COVID-19 Outbreak: A Cross-Sectional Study
A decline in the fertility rate has been observed worldwide, which hampers social development severely. Given the impacts of COVID-19 on individuals and society, it is of great significance to investigate the fertility intention of reproductive couples under COVID-19. The convenience sampling method was used to obtain our study sample. The self-administered questionnaire included the following components: sociodemographic characteristics (age, residence, education, occupation, characteristics of the couples, and annual household income), reproductive history (parity, number of children, child gender, and duration of preparing pregnancy), and attitudes toward COVID-19, was distributed online via an applet of WeChat. The results showed that among 4,133 valid questionnaires, 1,091 had fertility intention before COVID-19, whereas 3,042 did not, indicating a fertility intention rate of 26.4% among participating couples. Of the 1,091 couples who had fertility intention before COVID-19, 520 (47.7%) were affected by the outbreak, whereas 571 (52.3%) were not. By multivariable logistic regression analysis, we further found that couples living in Hubei Province, the epicenter in China (OR 2.20, 95% CI 1.35–3.60), and couples who prepared for pregnancy longer before COVID-19 (OR 1.19, 95% CI 1.06–1.33) were more likely to change their fertility intention under the pandemic. In addition, most of the participants reported their fertility intention was affected by the inconvenience of seeking medical service under COVID-19. Therefore, more forms of medical services to provide convenience for patients might be effective ways to reverse the declined fertility intention rate in facing COVID-19.
- 6 January 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cybernetics
- Vol. 53 (7) , 4079-4093
The Hot Strip Mill Scheduling Problem With Uncertainty: Robust Optimization Models and Solution Approaches
In this article, we focus on a biobjective hot strip mill (HSM) scheduling problem arising in the steel industry. Besides the conventional objective regarding penalty costs, we have also considered minimizing the total starting times of rolling operations in order to reduce the energy consumption for slab reheating. The problem is complicated by the inevitable uncertainty in rolling processing times, which means deterministic scheduling models will be ineffective. To obtain robust production schedules with satisfactory performance under all possible conditions, we apply the robust optimization (RO) approach to model and solve the scheduling problem. First, an RO model and an equivalent mixed-integer linear programming model are constructed to describe the HSM scheduling problem with uncertainty. Then, we devise an improved Benders’ decomposition algorithm to solve the RO model and obtain exactly optimal solutions. Next, for coping with large-sized instances, a multiobjective particle swarm optimization algorithm with an embedded local search strategy is proposed to handle the biobjective scheduling problem and find the set of Pareto-optimal solutions. Finally, we conduct extensive computational tests to verify the proposed algorithms. Results show that the exact algorithm is effective for relatively small instances and the metaheuristic algorithm can achieve satisfactory solution quality for both small- and large-sized instances of the problem.
- 6 January 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cybernetics
- Vol. 53 (7) , 4079-4093
The Hot Strip Mill Scheduling Problem With Uncertainty: Robust Optimization Models and Solution Approaches
In this article, we focus on a biobjective hot strip mill (HSM) scheduling problem arising in the steel industry. Besides the conventional objective regarding penalty costs, we have also considered minimizing the total starting times of rolling operations in order to reduce the energy consumption for slab reheating. The problem is complicated by the inevitable uncertainty in rolling processing times, which means deterministic scheduling models will be ineffective. To obtain robust production schedules with satisfactory performance under all possible conditions, we apply the robust optimization (RO) approach to model and solve the scheduling problem. First, an RO model and an equivalent mixed-integer linear programming model are constructed to describe the HSM scheduling problem with uncertainty. Then, we devise an improved Benders’ decomposition algorithm to solve the RO model and obtain exactly optimal solutions. Next, for coping with large-sized instances, a multiobjective particle swarm optimization algorithm with an embedded local search strategy is proposed to handle the biobjective scheduling problem and find the set of Pareto-optimal solutions. Finally, we conduct extensive computational tests to verify the proposed algorithms. Results show that the exact algorithm is effective for relatively small instances and the metaheuristic algorithm can achieve satisfactory solution quality for both small- and large-sized instances of the problem.
- 9 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 44 (12) , 10222-10235
Temporal-Spatial Causal Interpretations for Vision-Based Reinforcement Learning
Deep reinforcement learning (RL) agents are becoming increasingly proficient in a range of complex control tasks. However, the agent’s behavior is usually difficult to interpret due to the introduction of black-box function, making it difficult to acquire the trust of users. Although there have been some interesting interpretation methods for vision-based RL, most of them cannot uncover temporal causal information, raising questions about their reliability. To address this problem, we present a temporal-spatial causal interpretation (TSCI) model to understand the agent’s long-term behavior, which is essential for sequential decision-making. TSCI model builds on the formulation of temporal causality, which reflects the temporal causal relations between sequential observations and decisions of RL agent. Then a separate causal discovery network is employed to identify temporal-spatial causal features, which are constrained to satisfy the temporal causality. TSCI model is applicable to recurrent agents and can be used to discover causal features with high efficiency once trained. The empirical results show that TSCI model can produce high-resolution and sharp attention masks to highlight task-relevant temporal-spatial information that constitutes most evidence about how vision-based RL agents make sequential decisions. In addition, we further demonstrate that our method is able to provide valuable causal interpretations for vision-based RL agents from the temporal perspective.
- 9 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 44 (12) , 10222-10235
Temporal-Spatial Causal Interpretations for Vision-Based Reinforcement Learning
Deep reinforcement learning (RL) agents are becoming increasingly proficient in a range of complex control tasks. However, the agent’s behavior is usually difficult to interpret due to the introduction of black-box function, making it difficult to acquire the trust of users. Although there have been some interesting interpretation methods for vision-based RL, most of them cannot uncover temporal causal information, raising questions about their reliability. To address this problem, we present a temporal-spatial causal interpretation (TSCI) model to understand the agent’s long-term behavior, which is essential for sequential decision-making. TSCI model builds on the formulation of temporal causality, which reflects the temporal causal relations between sequential observations and decisions of RL agent. Then a separate causal discovery network is employed to identify temporal-spatial causal features, which are constrained to satisfy the temporal causality. TSCI model is applicable to recurrent agents and can be used to discover causal features with high efficiency once trained. The empirical results show that TSCI model can produce high-resolution and sharp attention masks to highlight task-relevant temporal-spatial information that constitutes most evidence about how vision-based RL agents make sequential decisions. In addition, we further demonstrate that our method is able to provide valuable causal interpretations for vision-based RL agents from the temporal perspective.
- 2 February 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 70 (00189456) , 1-10
Dual Attention-Based Temporal Convolutional Network for Fault Prognosis Under Time-Varying Operating Conditions
Fault prognosis under time-varying operating conditions is critical to predictive maintenance of complex industrial machines. However, prognostic methods proposed in the literature toward varying operating conditions typically focus on limited discrete states with clear boundaries, which are not always encountered in practical scenarios. This article aims to propose a more general method that can handle complex operating conditions with more nonlinearity and dynamics. In this work, a novel deep model named dual attention-based temporal convolutional network (DATCN), is developed to integrate multisensor data with time-varying operating settings for health assessment of industrial machines. An efficient deep temporal convolutional network is introduced to extract degradation features for sequence modeling. A novel dual attention model is developed to adaptively capture the temporal dynamics of sensor data and operating data, further enhancing the feature representations and highlighting the degradation information under time-varying operating settings. Besides, domain adaptation layer with multikernel maximum mean discrepancies (MK-MMD) is embedded to mitigate the domain-shift issue caused by varying operating conditions. The experimental results on the popular turbofan engine data set and a real-world Ion mill etching process validate the effectiveness of the proposed method and show its superiority in yielding significant improvements of prognostic accuracy relative to existing state-of-the-art methods.
- 2 February 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 70 (00189456) , 1-10
Dual Attention-Based Temporal Convolutional Network for Fault Prognosis Under Time-Varying Operating Conditions
Fault prognosis under time-varying operating conditions is critical to predictive maintenance of complex industrial machines. However, prognostic methods proposed in the literature toward varying operating conditions typically focus on limited discrete states with clear boundaries, which are not always encountered in practical scenarios. This article aims to propose a more general method that can handle complex operating conditions with more nonlinearity and dynamics. In this work, a novel deep model named dual attention-based temporal convolutional network (DATCN), is developed to integrate multisensor data with time-varying operating settings for health assessment of industrial machines. An efficient deep temporal convolutional network is introduced to extract degradation features for sequence modeling. A novel dual attention model is developed to adaptively capture the temporal dynamics of sensor data and operating data, further enhancing the feature representations and highlighting the degradation information under time-varying operating settings. Besides, domain adaptation layer with multikernel maximum mean discrepancies (MK-MMD) is embedded to mitigate the domain-shift issue caused by varying operating conditions. The experimental results on the popular turbofan engine data set and a real-world Ion mill etching process validate the effectiveness of the proposed method and show its superiority in yielding significant improvements of prognostic accuracy relative to existing state-of-the-art methods.
- 21 January 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 44 (01628828) , 1
Regularizing Deep Networks with Semantic Data Augmentation
Data augmentation is widely known as a simple yet surprisingly effective technique for regularizing deep networks. Conventional data augmentation schemes, e.g., flipping, translation or rotation, are low-level, data-independent and class-agnostic operations, leading to limited diversity for augmented samples. To this end, we propose a novel semantic data augmentation algorithm to complement traditional approaches. The proposed method is inspired by the intriguing property that deep networks are effective in learning linearized features, i.e., certain directions in the deep feature space correspond to meaningful semantic transformations, e.g., changing the background or view angle of an object. Based on this observation, translating training samples along many such directions in the feature space can effectively augment the dataset for more diversity. To implement this idea, we first introduce a sampling based method to obtain semantically meaningful directions efficiently. Then, an upper bound of the expected cross-entropy (CE) loss on the augmented training set is derived by assuming the number of augmented samples goes to infinity, yielding a highly efficient algorithm. In fact, we show that the proposed implicit semantic data augmentation (ISDA) algorithm amounts to minimizing a novel robust CE loss, which adds minimal extra computational cost to a normal training procedure. In addition to supervised learning, ISDA can be applied to semi-supervised learning tasks under the consistency regularization framework, where ISDA amounts to minimizing the upper bound of the expected KL-divergence between the augmented features and the original features. Although being simple, ISDA consistently improves the generalization performance of popular deep models (e.g., ResNets and DenseNets) on a variety of datasets, i.e., CIFAR-10, CIFAR-100, SVHN, ImageNet, and Cityscapes. Code for reproducing our results is available at https://github.com/blackfeather-wang/ISDA-for-Deep-Networks .
- 21 January 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 44 (01628828) , 1
Regularizing Deep Networks with Semantic Data Augmentation
Data augmentation is widely known as a simple yet surprisingly effective technique for regularizing deep networks. Conventional data augmentation schemes, e.g., flipping, translation or rotation, are low-level, data-independent and class-agnostic operations, leading to limited diversity for augmented samples. To this end, we propose a novel semantic data augmentation algorithm to complement traditional approaches. The proposed method is inspired by the intriguing property that deep networks are effective in learning linearized features, i.e., certain directions in the deep feature space correspond to meaningful semantic transformations, e.g., changing the background or view angle of an object. Based on this observation, translating training samples along many such directions in the feature space can effectively augment the dataset for more diversity. To implement this idea, we first introduce a sampling based method to obtain semantically meaningful directions efficiently. Then, an upper bound of the expected cross-entropy (CE) loss on the augmented training set is derived by assuming the number of augmented samples goes to infinity, yielding a highly efficient algorithm. In fact, we show that the proposed implicit semantic data augmentation (ISDA) algorithm amounts to minimizing a novel robust CE loss, which adds minimal extra computational cost to a normal training procedure. In addition to supervised learning, ISDA can be applied to semi-supervised learning tasks under the consistency regularization framework, where ISDA amounts to minimizing the upper bound of the expected KL-divergence between the augmented features and the original features. Although being simple, ISDA consistently improves the generalization performance of popular deep models (e.g., ResNets and DenseNets) on a variety of datasets, i.e., CIFAR-10, CIFAR-100, SVHN, ImageNet, and Cityscapes. Code for reproducing our results is available at https://github.com/blackfeather-wang/ISDA-for-Deep-Networks .
- Open Access
Tighter Bound Estimation of Sensitivity Analysis for Incremental and Decremental Data Modification
In large-scale classification problems, the data set always be faced with frequent updates when a part of the data is added to or removed from the original data set. In this case, conventional incremental learning, which updates an existing classifier by explicitly modeling the data modification, is more efficient than retraining a new classifier from scratch. However, sometimes, we are more interested in determining whether we should update the classifier or performing some sensitivity analysis tasks. To deal with these such tasks, we propose an algorithm to make rational inferences about the updated linear classifier without exactly updating the classifier. Specifically, the proposed algorithm can be used to estimate the upper and lower bounds of the updated classifier's coefficient matrix with a low computational complexity related to the size of the updated dataset. Both theoretical analysis and experiment results show that the proposed approach is superior to existing methods in terms of tightness of coefficients' bounds and computational complexity.
Open AccessTighter Bound Estimation of Sensitivity Analysis for Incremental and Decremental Data Modification
In large-scale classification problems, the data set always be faced with frequent updates when a part of the data is added to or removed from the original data set. In this case, conventional incremental learning, which updates an existing classifier by explicitly modeling the data modification, is more efficient than retraining a new classifier from scratch. However, sometimes, we are more interested in determining whether we should update the classifier or performing some sensitivity analysis tasks. To deal with these such tasks, we propose an algorithm to make rational inferences about the updated linear classifier without exactly updating the classifier. Specifically, the proposed algorithm can be used to estimate the upper and lower bounds of the updated classifier's coefficient matrix with a low computational complexity related to the size of the updated dataset. Both theoretical analysis and experiment results show that the proposed approach is superior to existing methods in terms of tightness of coefficients' bounds and computational complexity.
- 13 November 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 44 (01628828) , 1
Self-Supervised Discovering of Interpretable Features for Reinforcement Learning
Deep reinforcement learning (RL) has recently led to many breakthroughs on a range of complex control tasks. However, the agent’s decision-making process is generally not transparent. The lack of interpretability hinders the applicability of RL in safety-critical scenarios. While several methods have attempted to interpret vision-based RL, most come without detailed explanation for the agent’s behavior. In this paper, we propose a self-supervised interpretable framework, which can discover interpretable features to enable easy understanding of RL agents even for non-experts. Specifically, a self-supervised interpretable network (SSINet) is employed to produce fine-grained attention masks for highlighting task-relevant information, which constitutes most evidence for the agent’s decisions. We verify and evaluate our method on several Atari 2600 games as well as Duckietown, which is a challenging self-driving car simulator environment. The results show that our method renders empirical evidences about how the agent makes decisions and why the agent performs well or badly, especially when transferred to novel scenes. Overall, our method provides valuable insight into the internal decision-making process of vision-based RL. In addition, our method does not use any external labelled data, and thus demonstrates the possibility to learn high-quality mask through a self-supervised manner, which may shed light on new paradigms for label-free vision learning such as self-supervised segmentation and detection.
- 13 November 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 44 (01628828) , 1
Self-Supervised Discovering of Interpretable Features for Reinforcement Learning
Deep reinforcement learning (RL) has recently led to many breakthroughs on a range of complex control tasks. However, the agent’s decision-making process is generally not transparent. The lack of interpretability hinders the applicability of RL in safety-critical scenarios. While several methods have attempted to interpret vision-based RL, most come without detailed explanation for the agent’s behavior. In this paper, we propose a self-supervised interpretable framework, which can discover interpretable features to enable easy understanding of RL agents even for non-experts. Specifically, a self-supervised interpretable network (SSINet) is employed to produce fine-grained attention masks for highlighting task-relevant information, which constitutes most evidence for the agent’s decisions. We verify and evaluate our method on several Atari 2600 games as well as Duckietown, which is a challenging self-driving car simulator environment. The results show that our method renders empirical evidences about how the agent makes decisions and why the agent performs well or badly, especially when transferred to novel scenes. Overall, our method provides valuable insight into the internal decision-making process of vision-based RL. In addition, our method does not use any external labelled data, and thus demonstrates the possibility to learn high-quality mask through a self-supervised manner, which may shed light on new paradigms for label-free vision learning such as self-supervised segmentation and detection.
- 19 July 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 29 (10577149) , 199-213
A Graph Embedding Framework for Maximum Mean Discrepancy-Based Domain Adaptation Algorithms
Domain adaptation aims to deal with learning problems in which the labeled training data and unlabeled testing data are differently distributed. Maximum Mean Discrepancy (MMD), as a distribution distance measure, is minimized in various domain adaptation algorithms for eliminating domain divergence. We analyze empirical MMD from the point of view of graph embedding. It is discovered from the MMD intrinsic graph that, when the empirical MMD is minimized, the compactness within each domain and each class is simultaneously reduced. Therefore, points from different classes may mutually overlap, leading to unsatisfactory classification results. To deal with this issue, we present a graph embedding framework with intrinsic and penalty graphs for MMD-based domain adaptation algorithms. In the framework, we revise the intrinsic graph of MMD-based algorithms such that the within-class scatter is minimized, and thus the new features are discriminative. Two strategies are proposed. Based on the strategies, we instantiate the framework by exploiting four models. Each model has a penalty graph characterizing certain similarity property that should be avoided. Comprehensive experiments on visual cross-domain benchmark datasets demonstrate that the proposed models can greatly enhance the classification performance compared to state-of-the-art methods.
- 19 July 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 29 (10577149) , 199-213
A Graph Embedding Framework for Maximum Mean Discrepancy-Based Domain Adaptation Algorithms
Domain adaptation aims to deal with learning problems in which the labeled training data and unlabeled testing data are differently distributed. Maximum Mean Discrepancy (MMD), as a distribution distance measure, is minimized in various domain adaptation algorithms for eliminating domain divergence. We analyze empirical MMD from the point of view of graph embedding. It is discovered from the MMD intrinsic graph that, when the empirical MMD is minimized, the compactness within each domain and each class is simultaneously reduced. Therefore, points from different classes may mutually overlap, leading to unsatisfactory classification results. To deal with this issue, we present a graph embedding framework with intrinsic and penalty graphs for MMD-based domain adaptation algorithms. In the framework, we revise the intrinsic graph of MMD-based algorithms such that the within-class scatter is minimized, and thus the new features are discriminative. Two strategies are proposed. Based on the strategies, we instantiate the framework by exploiting four models. Each model has a penalty graph characterizing certain similarity property that should be avoided. Comprehensive experiments on visual cross-domain benchmark datasets demonstrate that the proposed models can greatly enhance the classification performance compared to state-of-the-art methods.