Yulong Wang

Yulong Wang, Tong Sun, Shenghong Li et al.

Adversarial attacks and defenses in machine learning and deep neural network have been gaining significant attention due to the rapidly growing applications of deep learning in the Internet and relevant scenarios. This survey provides a comprehensive overview of the recent advancements in the field of adversarial attack and defense techniques, with a focus on deep neural network-based classification models. Specifically, we conduct a comprehensive classification of recent adversarial attack methods and state-of-the-art adversarial defense techniques based on attack principles, and present them in visually appealing tables and tree diagrams. This is based on a rigorous evaluation of the existing works, including an analysis of their strengths and limitations. We also categorize the methods into counter-attack detection and robustness enhancement, with a specific focus on regularization-based methods for enhancing robustness. New avenues of attack are also explored, including search-based, decision-based, drop-based, and physical-world attacks, and a hierarchical classification of the latest defense methods is provided, highlighting the challenges of balancing training costs with performance, maintaining clean accuracy, overcoming the effect of gradient masking, and ensuring method transferability. At last, the lessons learned and open challenges are summarized with future research opportunities recommended.

Xuebin Zhao, Hong Chen, Yingjie Wang et al.

Recently, the scheme of model-X knockoffs was proposed as a promising solution to address controlled feature selection under high-dimensional finite-sample settings. However, the procedure of model-X knockoffs depends heavily on the coefficient-based feature importance and only concerns the control of false discovery rate (FDR). To further improve its adaptivity and flexibility, in this paper, we propose an error-based knockoff inference method by integrating the knockoff features, the error-based feature importance statistics, and the stepdown procedure together. The proposed inference procedure does not require specifying a regression model and can handle feature selection with theoretical guarantees on controlling false discovery proportion (FDP), FDR, or k-familywise error rate (k-FWER). Empirical evaluations demonstrate the competitive performance of our approach on both simulated and real data.

Ang Lv, Yuhan Chen, Kaiyi Zhang et al.

In this paper, we delve into several mechanisms employed by Transformer-based language models (LLMs) for factual recall tasks. We outline a pipeline consisting of three major steps: (1) Given a prompt ``The capital of France is,'' task-specific attention heads extract the topic token, such as ``France,'' from the context and pass it to subsequent MLPs. (2) As attention heads' outputs are aggregated with equal weight and added to the residual stream, the subsequent MLP acts as an ``activation,'' which either erases or amplifies the information originating from individual heads. As a result, the topic token ``France'' stands out in the residual stream. (3) A deep MLP takes ``France'' and generates a component that redirects the residual stream towards the direction of the correct answer, i.e., ``Paris.'' This procedure is akin to applying an implicit function such as ``get\_capital($X$),'' and the argument $X$ is the topic token information passed by attention heads. To achieve the above quantitative and qualitative analysis for MLPs, we proposed a novel analytic method aimed at decomposing the outputs of the MLP into components understandable by humans. Additionally, we observed a universal anti-overconfidence mechanism in the final layer of models, which suppresses correct predictions. We mitigate this suppression by leveraging our interpretation to improve factual recall confidence. The above interpretations are evaluated across diverse tasks spanning various domains of factual knowledge, using various language models from the GPT-2 families, 1.3B OPT, up to 7B Llama-2, and in both zero- and few-shot setups.

Yujie Lu, Long Wan, Nayu Ding et al.

Neural implicit representation of geometric shapes has witnessed considerable advancements in recent years. However, common distance field based implicit representations, specifically signed distance field (SDF) for watertight shapes or unsigned distance field (UDF) for arbitrary shapes, routinely suffer from degradation of reconstruction accuracy when converting to explicit surface points and meshes. In this paper, we introduce a novel neural implicit representation based on unsigned orthogonal distance fields (UODFs). In UODFs, the minimal unsigned distance from any spatial point to the shape surface is defined solely in one orthogonal direction, contrasting with the multi-directional determination made by SDF and UDF. Consequently, every point in the 3D UODFs can directly access its closest surface points along three orthogonal directions. This distinctive feature leverages the accurate reconstruction of surface points without interpolation errors. We verify the effectiveness of UODFs through a range of reconstruction examples, extending from simple watertight or non-watertight shapes to complex shapes that include hollows, internal or assembling structures.

Yuya Sasaki, Jing Tao, Yulong Wang

Motivated by the empirical observation of power-law distributions in the credits (e.g., "likes") of viral social media posts, we introduce a high-dimensional tail index regression model and propose methods for estimation and inference of its parameters. First, we present a regularized estimator, establish its consistency, and derive its convergence rate. Second, we introduce a debiasing technique for the regularized estimator to facilitate inference and prove its asymptotic normality. Third, we extend our approach to handle large-scale online streaming data using stochastic gradient descent. Simulation studies corroborate our theoretical findings. We apply these methods to the text analysis of viral posts on X (formerly Twitter) related to LGBTQ+ topics.

Yulong Wang, Tianxiang Li, Shenghong Li et al.

Deep Neural Networks (DNNs) are vulnerable to adversarial examples, while adversarial attack models, e.g., DeepFool, are on the rise and outrunning adversarial example detection techniques. This paper presents a new adversarial example detector that outperforms state-of-the-art detectors in identifying the latest adversarial attacks on image datasets. Specifically, we propose to use sentiment analysis for adversarial example detection, qualified by the progressively manifesting impact of an adversarial perturbation on the hidden-layer feature maps of a DNN under attack. Accordingly, we design a modularized embedding layer with the minimum learnable parameters to embed the hidden-layer feature maps into word vectors and assemble sentences ready for sentiment analysis. Extensive experiments demonstrate that the new detector consistently surpasses the state-of-the-art detection algorithms in detecting the latest attacks launched against ResNet and Inception neutral networks on the CIFAR-10, CIFAR-100 and SVHN datasets. The detector only has about 2 million parameters, and takes shorter than 4.6 milliseconds to detect an adversarial example generated by the latest attack models using a Tesla K80 GPU card.

Yulong Wang, Shenghong Li, Wei Ni et al.

Many location-based services use Received Signal Strength (RSS) measurements due to their universal availability. In this paper, we study the association of a large number of low-cost Internet-of-Things (IoT) sensors and their possible installation locations, which can enable various sensing and automation-related applications. We propose an efficient approach to solve the corresponding permutation combinatorial optimization problem, which integrates continuous space cooperative localization and permutation space likelihood ascent search. A convex relaxation-based optimization is designed to estimate the coarse locations of blindfolded devices in continuous 3D spaces, which are then projected to the feasible permutation space. An efficient Cramér-Rao Lower Bound based likelihood ascent search algorithm is proposed to refine the solution. Extensive experiments were conducted to evaluate the performance of the proposed approach, which show that the proposed approach significantly outperforms state-of-the-art combinatorial optimization algorithms and achieves close-to-100% accuracy with affordable execution time.

Yulong Wang, Yuan Yan Tang, Luoqing Li et al.

Representation based classification (RC) methods such as sparse RC (SRC) have shown great potential in face recognition in recent years. Most previous RC methods are based on the conventional regression models, such as lasso regression, ridge regression or group lasso regression. These regression models essentially impose a predefined assumption on the distribution of the noise variable in the query sample, such as the Gaussian or Laplacian distribution. However, the complicated noises in practice may violate the assumptions and impede the performance of these RC methods. In this paper, we propose a modal regression based atomic representation and classification (MRARC) framework to alleviate such limitation. Unlike previous RC methods, the MRARC framework does not require the noise variable to follow any specific predefined distributions. This gives rise to the capability of MRARC in handling various complex noises in reality. Using MRARC as a general platform, we also develop four novel RC methods for unimodal and multimodal face recognition, respectively. In addition, we devise a general optimization algorithm for the unified MRARC framework based on the alternating direction method of multipliers (ADMM) and half-quadratic theory. The experiments on real-world data validate the efficacy of MRARC for robust face recognition.