Maximum-Entropy Adversarial Data Augmentation for Improved Generalization and Robustness

Adversarial data augmentation has shown promise for training robust deep neural networks against unforeseen data shifts or corruptions. However, it is difficult to define heuristics to generate effective fictitious target distributions containing "hard" adversarial perturbations that are largely different from the source distribution. In this paper, we propose a novel and effective regularization term for adversarial data augmentation. We theoretically derive it from the information bottleneck principle, which results in a maximum-entropy formulation. Intuitively, this regularization term encourages perturbing the underlying source distribution to enlarge predictive uncertainty of the current model, so that the generated "hard" adversarial perturbations can improve the model robustness during training. Experimental results on three standard benchmarks demonstrate that our method consistently outperforms the existing state of the art by a statistically significant margin.

Contrastive Clustering

In this paper, we propose a one-stage online clustering method called Contrastive Clustering (CC) which explicitly performs the instance- and cluster-level contrastive learning. To be specific, for a given dataset, the positive and negative instance pairs are constructed through data augmentations and then projected into a feature space. Therein, the instance- and cluster-level contrastive learning are respectively conducted in the row and column space by maximizing the similarities of positive pairs while minimizing those of negative ones. Our key observation is that the rows of the feature matrix could be regarded as soft labels of instances, and accordingly the columns could be further regarded as cluster representations. By simultaneously optimizing the instance- and cluster-level contrastive loss, the model jointly learns representations and cluster assignments in an end-to-end manner. Extensive experimental results show that CC remarkably outperforms 17 competitive clustering methods on six challenging image benchmarks. In particular, CC achieves an NMI of 0.705 (0.431) on the CIFAR-10 (CIFAR-100) dataset, which is an up to 19\% (39\%) performance improvement compared with the best baseline.

Structured Graph Learning for Clustering and Semi-supervised Classification

Graphs have become increasingly popular in modeling structures and interactions in a wide variety of problems during the last decade. Graph-based clustering and semi-supervised classification techniques have shown impressive performance. This paper proposes a graph learning framework to preserve both the local and global structure of data. Specifically, our method uses the self-expressiveness of samples to capture the global structure and adaptive neighbor approach to respect the local structure. Furthermore, most existing graph-based methods conduct clustering and semi-supervised classification on the graph learned from the original data matrix, which doesn't have explicit cluster structure, thus they might not achieve the optimal performance. By considering rank constraint, the achieved graph will have exactly $c$ connected components if there are $c$ clusters or classes. As a byproduct of this, graph learning and label inference are jointly and iteratively implemented in a principled way. Theoretically, we show that our model is equivalent to a combination of kernel k-means and k-means methods under certain condition. Extensive experiments on clustering and semi-supervised classification demonstrate that the proposed method outperforms other state-of-the-art methods.

You Only Look Yourself: Unsupervised and Untrained Single Image Dehazing Neural Network

In this paper, we study two challenging and less-touched problems in single image dehazing, namely, how to make deep learning achieve image dehazing without training on the ground-truth clean image (unsupervised) and a image collection (untrained). An unsupervised neural network will avoid the intensive labor collection of hazy-clean image pairs, and an untrained model is a ``real'' single image dehazing approach which could remove haze based on only the observed hazy image itself and no extra images is used. Motivated by the layer disentanglement idea, we propose a novel method, called you only look yourself (\textbf{YOLY}) which could be one of the first unsupervised and untrained neural networks for image dehazing. In brief, YOLY employs three jointly subnetworks to separate the observed hazy image into several latent layers, \textit{i.e.}, scene radiance layer, transmission map layer, and atmospheric light layer. After that, these three layers are further composed to the hazy image in a self-supervised manner. Thanks to the unsupervised and untrained characteristics of YOLY, our method bypasses the conventional training paradigm of deep models on hazy-clean pairs or a large scale dataset, thus avoids the labor-intensive data collection and the domain shift issue. Besides, our method also provides an effective learning-based haze transfer solution thanks to its layer disentanglement mechanism. Extensive experiments show the promising performance of our method in image dehazing compared with 14 methods on four databases.

Heterogeneous Representation Learning: A Review

The real-world data usually exhibits heterogeneous properties such as modalities, views, or resources, which brings some unique challenges wherein the key is Heterogeneous Representation Learning (HRL) termed in this paper. This brief survey covers the topic of HRL, centered around several major learning settings and real-world applications. First of all, from the mathematical perspective, we present a unified learning framework which is able to model most existing learning settings with the heterogeneous inputs. After that, we conduct a comprehensive discussion on the HRL framework by reviewing some selected learning problems along with the mathematics perspectives, including multi-view learning, heterogeneous transfer learning, Learning using privileged information and heterogeneous multi-task learning. For each learning task, we also discuss some applications under these learning problems and instantiates the terms in the mathematical framework. Finally, we highlight the challenges that are less-touched in HRL and present future research directions. To the best of our knowledge, there is no such framework to unify these heterogeneous problems, and this survey would benefit the community.

Knowledge as Priors: Cross-Modal Knowledge Generalization for Datasets without Superior Knowledge

Cross-modal knowledge distillation deals with transferring knowledge from a model trained with superior modalities (Teacher) to another model trained with weak modalities (Student). Existing approaches require paired training examples exist in both modalities. However, accessing the data from superior modalities may not always be feasible. For example, in the case of 3D hand pose estimation, depth maps, point clouds, or stereo images usually capture better hand structures than RGB images, but most of them are expensive to be collected. In this paper, we propose a novel scheme to train the Student in a Target dataset where the Teacher is unavailable. Our key idea is to generalize the distilled cross-modal knowledge learned from a Source dataset, which contains paired examples from both modalities, to the Target dataset by modeling knowledge as priors on parameters of the Student. We name our method "Cross-Modal Knowledge Generalization" and demonstrate that our scheme results in competitive performance for 3D hand pose estimation on standard benchmark datasets.

Learning to Learn Single Domain Generalization

We are concerned with a worst-case scenario in model generalization, in the sense that a model aims to perform well on many unseen domains while there is only one single domain available for training. We propose a new method named adversarial domain augmentation to solve this Out-of-Distribution (OOD) generalization problem. The key idea is to leverage adversarial training to create "fictitious" yet "challenging" populations, from which a model can learn to generalize with theoretical guarantees. To facilitate fast and desirable domain augmentation, we cast the model training in a meta-learning scheme and use a Wasserstein Auto-Encoder (WAE) to relax the widely used worst-case constraint. Detailed theoretical analysis is provided to testify our formulation, while extensive experiments on multiple benchmark datasets indicate its superior performance in tackling single domain generalization.

Learning Graph Embedding with Limited Labeled Data: An Efficient Sampling Approach

Semi-supervised graph embedding methods represented by graph convolutional network has become one of the most popular methods for utilizing deep learning approaches to process the graph-based data for applications. Mostly existing work focus on designing novel algorithm structure to improve the performance, but ignore one common training problem, i.e., could these methods achieve the same performance with limited labelled data? To tackle this research gap, we propose a sampling-based training framework for semi-supervised graph embedding methods to achieve better performance with smaller training data set. The key idea is to integrate the sampling theory and embedding methods by a pipeline form, which has the following advantages: 1) the sampled training data can maintain more accurate graph characteristics than uniformly chosen data, which eliminates the model deviation; 2) the smaller scale of training data is beneficial to reduce the human resource cost to label them; The extensive experiments show that the sampling-based method can achieve the same performance only with 10$\%$-50$\%$ of the scale of training data. It verifies that the framework could extend the existing semi-supervised methods to the scenarios with the extremely small scale of labelled data.

Improving Distant Supervised Relation Extraction by Dynamic Neural Network

Distant Supervised Relation Extraction (DSRE) is usually formulated as a problem of classifying a bag of sentences that contain two query entities, into the predefined relation classes. Most existing methods consider those relation classes as distinct semantic categories while ignoring their potential connection to query entities. In this paper, we propose to leverage this connection to improve the relation extraction accuracy. Our key ideas are twofold: (1) For sentences belonging to the same relation class, the expression style, i.e. words choice, can vary according to the query entities. To account for this style shift, the model should adjust its parameters in accordance with entity types. (2) Some relation classes are semantically similar, and the entity types appear in one relation may also appear in others. Therefore, it can be trained cross different relation classes and further enhance those classes with few samples, i.e., long-tail classes. To unify these two arguments, we developed a novel Dynamic Neural Network for Relation Extraction (DNNRE). The network adopts a novel dynamic parameter generator that dynamically generates the network parameters according to the query entity types and relation classes. By using this mechanism, the network can simultaneously handle the style shift problem and enhance the prediction accuracy for long-tail classes. Through our experimental study, we demonstrate the effectiveness of the proposed method and show that it can achieve superior performance over the state-of-the-art methods.

Scalable Global Alignment Graph Kernel Using Random Features: From Node Embedding to Graph Embedding

Graph kernels are widely used for measuring the similarity between graphs. Many existing graph kernels, which focus on local patterns within graphs rather than their global properties, suffer from significant structure information loss when representing graphs. Some recent global graph kernels, which utilizes the alignment of geometric node embeddings of graphs, yield state-of-the-art performance. However, these graph kernels are not necessarily positive-definite. More importantly, computing the graph kernel matrix will have at least quadratic {time} complexity in terms of the number and the size of the graphs. In this paper, we propose a new family of global alignment graph kernels, which take into account the global properties of graphs by using geometric node embeddings and an associated node transportation based on earth mover's distance. Compared to existing global kernels, the proposed kernel is positive-definite. Our graph kernel is obtained by defining a distribution over \emph{random graphs}, which can naturally yield random feature approximations. The random feature approximations lead to our graph embeddings, which is named as "random graph embeddings" (RGE). In particular, RGE is shown to achieve \emph{(quasi-)linear scalability} with respect to the number and the size of the graphs. The experimental results on nine benchmark datasets demonstrate that RGE outperforms or matches twelve state-of-the-art graph classification algorithms.