Search

In this thesis, we use information-theoretic measures to quantify the caused disruption and the probability of detection of the data injection attacks. Specifically the attacker minimizes the mutual information between the state variables and the compromised measurements to minimize the amount of information acquired by the operator from the measurements about the state variables. Also the attacker minimizes the Kullback-Leibler divergence between the distribution of measurements with attack and without attack to minimize the probability of detection. The stealth attacks achieve these two contradictive objectives by minimizing the equal sum of them, which is generalized to the weighted sum later. Closed-form expression for the optimal Gaussian attack is proposed for the stealth attacks and the generalized stealth attacks when the attacker prioritizes the probability of detection over the disruption. Additionally, a closed-form expression of the probability of detection is obtained. To inform the design guidelines for the corresponding weighting parameter, a concentration inequality upper bound is proposed for the probability of detection. RMT tools are used to characterize the ergodic performance of the attacks when the attacker only gets access to a limited number of samples of the state variables. For the non-asymptotic scenario, an upper bound is proposed for the ergodic performance, for which a simple convex optimization needs to be solved to compute it. For the asymptotic case, a closed-form expression is provided for the ergodic performance of the attacks.

A long standing goal of artificial intelligence is to enable machines to perceive the visual world and interact with humans using natural language. To achieve this goal, many computer vision and natural language processing techniques have been proposed during the past decades, especially deep convolutional neural networks (CNNs). However, most previous work mainly focus on the two sides separately, and few work have been done by connecting the vision and language modalities. Hence, the semantic gap between the two modalities still exists. To solve this, the overall objective of my PhD research is to design machine learning algorithms for visual content understanding by connecting the vision and language modalities. Towards this goal, we have developed several deep learning models combined with natural language processing techniques to represent and analyze the image/video and text data. We focus on a series of applications to demonstrate the effectiveness of the proposed models and obtain promising results. Firstly we show that tag-based image annotations exhibit many limitations for visual content representation, and then develop techniques to discover visual themes as an alternative by re-organizing the original image and tag set into a group of visual themes. More concretely, we extract visual feature and semantic feature from two trained deep learning models respectively, and then design a method to effectively evaluate the similarity between a pair of tags both visually and semantically. Next we cluster these tags into a set of visual themes based on their joint similarities. We conduct human-based evaluation and machine-based evaluation to demonstrate the usefulness and rationality of the discovered visual themes and indicate their potential usage for automatically managing user photos. Secondly we develop a novel framework for understanding complex video scenes involving objects, humans, scene backgrounds and the interactions between them. In this work, we propose to automatically discover semantic information for these videos from an unsupervised manner. We do this by introducing a set of semantic attributes derived from a joint image and text corpora. Then we re-train a deep constitutional neural network to produce visual and semantic features simultaneously. Our trained model encodes the complex information including the whole visual scene, object appearance/property and motion. We apply our model to solve the video summarization problem by adopting a partially near duplicate image discovery technique to cluster visually and semantically consistent video frames together. The experimental results demonstrate the effectiveness of the semantic attributes in assisting the visual features in the video summarization and our new technique (SASUM) achieves state-of-the-art performance. Thirdly we decide to depict and interpret the traffic scene using vehicle objects. Firstly, we collect an hour-long traffic video with a resolution of 3840X2160 at 5 busy intersections of a megacity by flying an UAV during the rush hours. We then build a UavCT dataset containing over 64K annotated vehicles in 17K 512X512 images. In the next, we design and train a deep constitutional neural network from scratch to detect and localize different types of road vehicles (i.e. car, bus and truck), and propose a fast online tracking method to track and count vehicles in consecutive video frames. We design vehicle counting experiments on both image and video data to demonstrate the effectiveness of the proposed method. Fourthly we extend our work in the previous stage and explore more potential applications of our deep model for traffic scene understanding. Specifically, we first track all the target vehicles from the original video, and then recognize vehicle behaviors based on nearest neighbor search (clustering) and bidirectional long short-term memory (classification). By conducting comparative studies, we further demonstrate the effectiveness and versatility of our approach for object-based visual scene understanding. As last, we focus on TV series video understanding and develop techniques to recognize characters in these videos. Using label-level supervision, we transform the problem to multi-label classification and design a novel semantic projection network (SPNet) consisting of two stacked subnetworks with specially designed constraints. The first subnetwork is aiming to reconstruct the input feature activations from a trained single-label CNN, and the other one functions as a multi-label classifier which predicts the character labels as well as reconstructing the input visual features from the mapped semantic label space. We show such kind of mutual projection significantly benefits the character recognition by conducting experiments on three popular TV series video datasets. We also show that region-based prediction strategy could further improve the overall performance. Our work contribute to a developing research field that demonstrates the power of deep learning techniques in solving different visual recognition problems by connecting the vision and language modality and advance the state-of-the-art on various tasks.

Intrinsic drawbacks of Si, a well adapted material in photovoltaics, including the high valence band position compared to the proton reduction level, the unfavorable self-oxidation level compared to the one of water oxidation, and indirect energy band, prevent Si as a good candidate for applications in photoelectrochemical water splitting, despite its small band gap which matches the solar irradiance spectrum. Strategies need to be developed to overcome these problems. This dissertation presents a study of metal oxides integrated Si photoelectrode for solar driven water splitting. Different metal oxides including zinc oxide, titanium dioxide, indium tin oxide, and nickel oxide are coated on polished and nanotextured p -type and n-type Si from various methods. This heterogeneous uniform coating effectively replaces the problematic water Si interface which can be corrosive and oxidative to Si. The Schottky junction formed either using an n-type coating (ZnO and TiO₂) on p-type Si or a p-type coating (NiO) on n-type Si is able to effectively separate the photo-generated charges in Si. In addition, nanotextured Si is able to dramatically improve light absorption and provide enlarged surface chemical reaction sites. In this case, integration of metal oxides in the form of nanoscale structures on the nanostructured Si, a so-called branched nanowire structure or a "nanotree" array, is able to enhance the light absorption and reduce the radius of curvature to ease the gas evolution. Earth abundant transition metal, primarily Ni, was used to catalyze the water reduction and oxidation reaction. Solution-casted NiO from a cost-effective approach provides higher activity compared to the one prepared from gas phase sputtering. Surfactant and thermal treatment effect are studied to achieve an optimized catalytic activity. This catalyst is then successfully demonstrated on a surface inverted n-Si photoanode protected with a novel semitransparent conducting oxide. Considerable contributions have been made through this dissertation study. Suggestions on future improvements are provided