Regression Based Clustering by Deep Adversarial Learning

Learning an effective distance measurement between sample pairs plays an important role in visual analysis, where the training procedure largely relies on hard negative samples. However, hard negative samples usually account for the tiny minority in the training set, which may fail to fully describe the data distribution close to the decision boundary. In this paper, we present a deep adversarial metric learning (DAML) framework to generate synthetic hard negatives from the original negative samples, which is widely applicable to existing supervised deep metric learning algorithms. Different from existing sampling strategies which simply ignore numerous easy negatives, our DAML aim to exploit them by generating synthetic hard negatives adversarial to the learned metric as complements. We simultaneously train the feature embedding and hard negative generator in an adversarial manner, so that adequate and targeted synthetic hard negatives are created to learn more precise distance metrics. As a single transformation may not be powerful enough to describe the global input space under the attack of the hard negative generator, we further propose a deep adversarial multi-metric learning (DAMML) method by learning multiple local transformations for more complete description. We simultaneously exploit the collaborative and competitive relationships among multiple metrics, where the metrics display unity against the generator for effective distance measurement as well as compete for more training data through a metric discriminator to avoid overlapping. Extensive experimental results on five benchmark datasets show that our DAML and DAMML effectively boost the performance of existing deep metric learning approaches through adversarial learning.

Deep learning methods for person identification based on electroencephalographic (EEG) brain activity encounters the problem of exploiting the temporally correlated structures or recording session specific variability within EEG. Furthermore, recent methods have mostly trained and evaluated based on single session EEG data. We address this problem from an invariant representation learning perspective. We propose an adversarial inference approach to extend such deep learning models to learn session-invariant person-discriminative representations that can provide robustness in terms of longitudinal usability. Using adversarial learning within a deep convolutional network, we empirically assess and show improvements with our approach based on longitudinally collected EEG data for person identification from half-second EEG epochs.

The design of a new therapeutic agent is a time-consuming and expensive process. The rise of machine intelligence provides a grand opportunity of expeditiously discovering novel drug candidates through smart search in the vast molecular structural space. In this paper, we propose a new approach called adversarial deep evolutionary learning (ADEL) to search for novel molecules in the latent space of an adversarial generative model and keep improving the latent representation space. In ADEL, a custom-made adversarial autoencoder (AAE) model is developed and trained under a deep evolutionary learning (DEL) process. This involves an initial training of the AAE model, followed by an integration of multi-objective evolutionary optimization in the continuous latent representation space of the AAE rather than the discrete structural space of molecules. By using the AAE, an arbitrary distribution can be provided to the training of AAE such that the latent representation space is set to that distribution. This allows for a starting latent space from which new samples can be produced. Throughout the process of learning, new samples of high-quality are generated after each iteration of training and then added back into the full dataset. Therefore, allowing for a more comprehensive procedure of understanding the data structure. This combination of evolving data and continuous learning not only enables improvement in the generative model, but the data as well. By comparing ADEL to the previous work in DEL, we see that ADEL can obtain better property distributions.

Mobile user profiles are a summary of characteristics of user-specific mobile activities. Mobile user profiling is to extract a user's interest and behavioral patterns from mobile behavioral data. While some efforts have been made for mobile user profiling, existing methods can be improved via representation learning with awareness of substructures in users' behavioral graphs. Specifically, in this paper, we study the problem of mobile users profiling with POI check-in data. To this end, we first construct a graph, where a vertex is a POI category and an edge is the transition frequency of a user between two POI categories, to represent each user. We then formulate mobile user profiling as a task of representation learning from user behavioral graphs. We later develop a deep adversarial substructured learning framework for the task. This framework has two mutually-enhanced components. The first component is to preserve the structure of the entire graph, which is formulated as an encoding-decoding paradigm. In particular, the structure of the entire graph is preserved by minimizing reconstruction loss between an original graph and a reconstructed graph. The second component is to preserve the structure of subgraphs, which is formulated as a substructure detector based adversarial training paradigm. In particular, this paradigm includes a substructure detector and an adversarial trainer. Instead of using non-differentiable substructure detection algorithms, we pre-train a differentiable convolutional neural network as the detector to approximate these detection algorithms. The adversarial trainer is to match the detected substructure of the reconstructed graph to the detected substructure of the original graph. Also, we provide an effective solution for the optimization problems. Moreover, we exploit the learned representations of users for the next activity type prediction. Finally, we present extensive experimental results to demonstrate the improved performances of the proposed method.

Data-driven deep learning tasks for security related applications are gaining increasing popularity and achieving impressive performances. This thesis investigates adversarial vulnerabilities of such tasks in order to establish secure and reliable machine learning systems. Adversary attacks aim to extract private data from a model of a task and misguide the model so it yields wrong results or an answer desired by the attacker. This thesis studies potential adversarial attacks that may affect an existing deep learning model of a specific task. Novel approaches that expose security vulnerabilities of four typical deep learning models in three dominant tasks (i.e., matching, classification and regression) are developed. These models include image hashing for image authentication and retrieval, fake face imagery forensic detection, image classification and single object tracking. In the first model, image hashing converts images into codes that are supposed to be non-invertible. However, we prove that this can pose image privacy concerns, and propose two deep learning de-hashing neural networks to show that we can obtain high quality images that are inverted from given image hashes. In the second model, we address fake face image detection. Fake images that can escape an adversarial attacked detector are usually degraded versions of original images. We analyze the visual degradation in such face images, and show how to design attacks that result in visually imperceptible adversarial images. For the image classification model, instead of the conventionally employed visual distortion metric, we propose the use of perceptual models as a novel measure for adversarial example generation. We then propose two sets of attack methods that can generally be incorporated into all existing gradient-based attacks. Lastly, for the single object tracking model, we propose the concept of universally and physically feasible attacks on visual object tracking in real-world settings. We develop a novel attack framework and experimentally demonstrate the feasibility of the proposed concept. The adversarial explorations and examples provided in this thesis show how existing deep learning tasks and their models could be vulnerable to malicious attacks. This would help researchers design more secure and trustworthy models for digital media security and forensics.

Adversarial deep reinforcement learning based robust depth tracking control for underactuated autonomous underwater vehicle

Increasing numbers of software vulnerabilities are discovered every year whether they are reported publicly or discovered internally in proprietary code. These vulnerabilities can pose serious risk of exploit and result in system compromise, information leaks, or denial of service. We leveraged the wealth of C and C++ open-source code available to develop a largescale function-level vulnerability detection system using machine learning. To supplement existing labeled vulnerability datasets, we compiled a vast dataset of millions of open-source functions and labeled it with carefully-selected findings from three different static analyzers that indicate potential exploits. Using these datasets, we developed a fast and scalable vulnerability detection tool based on deep feature representation learning that directly interprets lexed source code. We evaluated our tool on code from both real software packages and the NIST SATE IV benchmark dataset. Our results demonstrate that deep feature representation learning on source code is a promising approach for automated software vulnerability detection.

Conversational agents are getting more popular and applied in a wide range of practical application areas. The main task of these agents is not only to generate context-appropriate responses to a given query but also to make the conversation human-like. Thanks to the ability of deep learning based models in natural language modeling, recent studies have made progress in designing conversational agents that can provide more semantically accurate responses. However, the naturalness in such conversation setting has not been given adequate attention in these studies. This paper aims to incorporate both important criteria of accuracy and naturalness of conversation in developing a new model for conversational agents. To this end, inspired by the idea of Turing test and the idea of adversarial learning strategy, we propose to design a model based on generative deep neural networks that interestingly allow to generate accurate responses optimized by the mechanics of imitating human-generated conversations. Experimental results demonstrate that the proposed models produce more natural and accurate responses, yielding significant gains in BLEU scores.

Advances in deep learning and representation learning have transformed item factor analysis (IFA) in the item response theory (IRT) literature by enabling more efficient and accurate parameter estimation. Variational autoencoders (VAEs) are widely used to model high-dimensional latent variables in this context, but the limited expressiveness of their inference networks can still hinder performance. We introduce adversarial variational Bayes (AVB) and an importance-weighted extension (IWAVB) as more flexible inference algorithms for IFA. By combining VAEs with generative adversarial networks (GANs), AVB uses an auxiliary discriminator network to frame estimation as a two-player game and removes the restrictive standard normal assumption on the latent variables. Theoretically, AVB and IWAVB can achieve likelihoods that match or exceed those of VAEs and importance-weighted autoencoders (IWAEs). In exploratory analyses of empirical data, IWAVB attained higher likelihoods than IWAE, indicating greater expressiveness. In confirmatory simulations, IWAVB achieved comparable mean-square error in parameter recovery while consistently yielding higher likelihoods, and it clearly outperformed IWAE when the latent distribution was multimodal. These findings suggest that IWAVB can scale IFA to complex, large-scale, and potentially multimodal settings, supporting closer integration of psychometrics with modern multimodal data analysis.

Although multi-view learning has made significant progress over the past few decades, it is still challenging due to the difficulty in modeling complex correlations among different views, especially under the context of view missing. To address the challenge, we propose a novel framework termed Cross Partial Multi-View Networks (CPM-Nets), which aims to fully and flexibly take advantage of multiple partial views. We first provide a formal definition of completeness and versatility for multi-view representation and then theoretically prove the versatility of the learned latent representations. For completeness, the task of learning latent multi-view representation is specifically translated to a degradation process by mimicking data transmission, such that the optimal tradeoff between consistency and complementarity across different views can be implicitly achieved. Equipped with adversarial strategy, our model stably imputes missing views, encoding information from all views for each sample to be encoded into latent representation to further enhance the completeness. Furthermore, a nonparametric classification loss is introduced to produce structured representations and prevent overfitting, which endows the algorithm with promising generalization under view-missing cases. Extensive experimental results validate the effectiveness of our algorithm over existing state of the arts for classification, representation learning and data imputation.

We study two important concepts in adversarial deep learning---adversarial training and generative adversarial network (GAN). Adversarial training is the technique used to improve the robustness of discriminator by combining adversarial attacker and discriminator in the training phase. GAN is commonly used for image generation by jointly optimizing discriminator and generator. We show these two concepts are indeed closely related and can be used to strengthen each other---adding a generator to the adversarial training procedure can improve the robustness of discriminators, and adding an adversarial attack to GAN training can improve the convergence speed and lead to better generators. Combining these two insights, we develop a framework called Rob-GAN to jointly optimize generator and discriminator in the presence of adversarial attacks---the generator generates fake images to fool discriminator; the adversarial attacker perturbs real images to fool the discriminator, and the discriminator wants to minimize loss under fake and adversarial images. Through this end-to-end training procedure, we are able to simultaneously improve the convergence speed of GAN training, the quality of synthetic images, and the robustness of discriminator under strong adversarial attacks. Experimental results demonstrate that the obtained classifier is more robust than the state-of-the-art adversarial training approach, and the generator outperforms SN-GAN on ImageNet-143.

Despite the continuous evolution and significant improvement of cybersecurity mechanisms, malware threats remain one of the most important concerns in cyberspace. Meanwhile, Android malware plays a big role in these ever-growing threats. In recent years, deep learning has become the dominant machine learning technique for malware detection and continues to make outstanding achievements. Deep graph representation learning is the task of embedding graph-structured data into a low-dimensional space using deep learning models. Recently, autoencoders have proven to be an effective way for deep representation learning. However, it is not straightforward to apply the idea of autoencoder to graph-structured data because of their irregular structure. In this paper, we present DroidMalGNN, a novel deep learning technique that combines autoencoders with graph neural networks (GNNs) to detect Android malware in an end-to-end manner. DroidMalGNN represents each Android application with an attributed function call graph (AFCG) that allows it to model complex relationships between data. For more efficiency, DroidMalGNN performs graph representation learning in a supervised manner where two autoencoders are trained with benign and malicious AFCGs separately. In this way, it generates two informative embedding vectors for each AFCG in a low-dimensional space and feeds them into a dense neural network to classify the AFCG as benign or malicious. Our experimental results show that DroidMalGNN can achieve good detection performance in terms of different evaluation measures.

Centroids-guided deep multi-view K-means clustering

A Comprehensive Survey on Deep Graph Representation Learning

Image cyberbullying detection and recognition using transfer deep machine learning