The escalating threat of phishing emails has become increasingly sophisticated with the rise of Large Language Models (LLMs). As attackers exploit LLMs to craft more convincing and evasive phishing emails, it is crucial to assess the resilience of current phishing defenses. In this study we conduct a comprehensive evaluation of traditional phishing detectors, such as Gmail Spam Filter, Apache SpamAssassin, and Proofpoint, as well as machine learning models like SVM, Logistic Regression, and Naive Bayes, in identifying both traditional and LLM-rephrased phishing emails. We also explore the emerging role of LLMs as phishing detection tools, a method already adopted by companies like NTT Security Holdings and JPMorgan Chase. Our results reveal notable declines in detection accuracy for rephrased emails across all detectors, highlighting critical weaknesses in current phishing defenses. As the threat landscape evolves, our findings underscore the need for stronger security controls and regulatory oversight on LLM-generated content to prevent its misuse in creating advanced phishing attacks. This study contributes to the development of more effective Cyber Threat Intelligence (CTI) by leveraging LLMs to generate diverse phishing variants that can be used for data augmentation, harnessing the power of LLMs to enhance phishing detection, and paving the way for more robust and adaptable threat detection systems.

Large language models (LLMs) require vast and high-quality training datasets, which are often difficult to annotate, particularly in domains such as finance. In finance, the datasets not only encompass large volumes of textual data but also involve intricate numerical reasoning, frequently presented in the form of tables. One significant challenge in this domain is the processing of exceptionally long financial statements and complex tables, which can hinder the reasoning capacity of LLMs and negatively impact their accuracy. To mitigate the adverse effects of lengthy text passages and complex tables, this study focuses on identifying and extracting key information (gold facts) that are essential for solving financial problems, with the aim of enhancing both reasoning quality and data efficiency. We investigate the disparities between various financial datasets and examine different model extraction techniques. Our experimental findings demonstrate that the precise extraction of these key facts from both textual and tabular data substantially improves the accuracy of LLMs-based question answering, while simultaneously reducing the amount of data necessary for effective model performance.

Language models (LMs) have demonstrated superior performance in detecting fraudulent activities on Blockchains. Nonetheless, the sheer volume of Blockchain data results in excessive memory and computational costs when training LMs from scratch, limiting their capabilities to large-scale applications. In this paper, we present ZipZap, a framework tailored to achieve both parameter and computational efficiency when training LMs on large-scale transaction data. First, with the frequency-aware compression, an LM can be compressed down to a mere 7.5% of its initial size with an imperceptible performance dip. This technique correlates the embedding dimension of an address with its occurrence frequency in the dataset, motivated by the observation that embeddings of low-frequency addresses are insufficiently trained and thus negating the need for a uniformly large dimension for knowledge representation. Second, ZipZap accelerates the speed through the asymmetric training paradigm: It performs transaction dropping and cross-layer parameter-sharing to expedite the pre-training process, while revert to the standard training paradigm for fine-tuning to strike a balance between efficiency and efficacy, motivated by the observation that the optimization goals of pre-training and fine-tuning are inconsistent. Evaluations on real-world, large-scale datasets demonstrate that ZipZap delivers notable parameter and computational efficiency improvements for training LMs. Our implementation is available at: https://github.com/git-disl/ZipZap.

Deep neural network ensembles combine the wisdom of multiple deep neural networks to improve the generalizability and robustness over individual networks. It has gained increasing popularity to study and apply deep ensemble techniques in the deep learning community. Some mission-critical applications utilize a large number of deep neural networks to form deep ensembles to achieve desired accuracy and resilience, which introduces high time and space costs for ensemble execution. However, it still remains a critical challenge whether a small subset of the entire deep ensemble can achieve the same or better generalizability and how to effectively identify these small deep ensembles for improving the space and time efficiency of ensemble execution. This article presents a novel deep ensemble pruning approach, which can efficiently identify smaller deep ensembles and provide higher ensemble accuracy than the entire deep ensemble of a large number of member networks. Our hierarchical ensemble pruning approach (HQ) leverages three novel ensemble pruning techniques. First, we show that the focal ensemble diversity metrics can accurately capture the complementary capacity of the member networks of an ensemble team, which can guide ensemble pruning. Second, we design a focal ensemble diversity based hierarchical pruning approach, which will iteratively find high quality deep ensembles with low cost and high accuracy. Third, we develop a focal diversity consensus method to integrate multiple focal diversity metrics to refine ensemble pruning results, where smaller deep ensembles can be effectively identified to offer high accuracy, high robustness and high ensemble execution efficiency. Evaluated using popular benchmark datasets, we demonstrate that the proposed hierarchical ensemble pruning approach can effectively identify high quality deep ensembles with better classification generalizability while being more time and space efficient in ensemble decision making. We have released the source codes on GitHub at https://github.com/git-disl/HQ-Ensemble .

Graph Neural Networks (GNNs) have enjoyed wide spread applications in graph-structured data. However, existing graph based applications commonly lack annotated data. GNNs are required to learn latent patterns from a limited amount of training data to perform inferences on a vast amount of test data. The increased complexity of GNNs, as well as a single point of model parameter initialization, usually lead to overfitting and sub-optimal performance. In addition, it is known that GNNs are vulnerable to adversarial attacks. In this paper, we push one step forward on the ensemble learning of GNNs with improved accuracy, generalization, and adversarial robustness. Following the principles of stochastic modeling, we propose a new method called GNN-Ensemble to construct an ensemble of random decision graph neural networks whose capacity can be arbitrarily expanded for improvement in performance. The essence of the method is to build multiple GNNs in randomly selected substructures in the topological space and subfeatures in the feature space, and then combine them for final decision making. These GNNs in different substructure and subfeature spaces generalize their classification in complementary ways. Consequently, their combined classification performance can be improved and overfitting on the training data can be effectively reduced. In the meantime, we show that GNN-Ensemble can significantly enhance the adversarial robustness against attacks on GNNs.

Deep neural network ensembles hold the potential of improving generalization performance for complex learning tasks. This paper presents formal analysis and empirical evaluation to show that heterogeneous deep ensembles with high ensemble diversity can effectively leverage model learning heterogeneity to boost ensemble robustness. We first show that heterogeneous DNN models trained for solving the same learning problem, e.g., object detection, can significantly strengthen the mean average precision (mAP) through our weighted bounding box ensemble consensus method. Second, we further compose ensembles of heterogeneous models for solving different learning problems, e.g., object detection and semantic segmentation, by introducing the connected component labeling (CCL) based alignment. We show that this two-tier heterogeneity driven ensemble construction method can compose an ensemble team that promotes high ensemble diversity and low negative correlation among member models of the ensemble, strengthening ensemble robustness against both negative examples and adversarial attacks. Third, we provide a formal analysis of the ensemble robustness in terms of negative correlation. Extensive experiments validate the enhanced robustness of heterogeneous ensembles in both benign and adversarial settings. The appendix and source codes are available on GitHub at https://github.com/git-disl/HeteRobust.

Gradient leakage attacks are dominating privacy threats in federated learning, despite the default privacy that training data resides locally at the clients. Differential privacy has been the de facto standard for privacy protection and is deployed in federated learning to mitigate privacy risks. However, much existing literature points out that differential privacy fails to defend against gradient leakage. The paper presents ModelCloak, a principled approach based on differential privacy noise, aiming for safe-sharing client local model updates. The paper is organized into three major components. First, we introduce the gradient leakage robustness trade-off, in search of the best balance between accuracy and leakage prevention. The trade-off relation is developed based on the behavior of gradient leakage attacks throughout the federated training process. Second, we demonstrate that a proper amount of differential privacy noise can offer the best accuracy performance within the privacy requirement under a fixed differential privacy noise setting. Third, we propose dynamic differential privacy noise and show that the privacy-utility trade-off can be further optimized with dynamic model perturbation, ensuring privacy protection, competitive accuracy, and leakage attack prevention simultaneously.

The increased digitization of commercial and consumer workflows, accelerated cloud adoption, and the growing sophistication of cyber criminals have resulted in ransomware emerging as a major threat of cyber attacks in cloud and data services. While research in malware detection can be partially adapted for ransomware, specific ransomware infection patterns can be leveraged to improve the detection efficiency. In this paper, we focus on identifying vulnerable targets in ransomware attacks, aiming to accelerate the ransomware detection process as well as enable better data backup policies design. Specifically, we make three contributions. First, we characterize lexical features and hierarchical file structure features on those ransomware infected files and folders. Second, we model the data backup as an attributed tree graph, learn a new feature representation of the nodes with graph neural networks, and train a classifier based on the new features. Third, with real-world snapshot backup instances, we demonstrate the superior performance of the graph representation learning based approach over several baselines. Compared to the traditional full-scan approaches, our finding suggests that the vulnerable ransomware attack target identification can result in a more efficient ransomware detection process via focused inspection of most vulnerable data in the backups. Our method can also be easily integrated into existing ransomware detection systems for accelerated cyber resiliency.

Deep neural networks (DNNs) have demonstrated impressive performance on many challenging machine learning tasks. However, DNNs are vulnerable to adversarial inputs generated by adding maliciously crafted perturbations to the benign inputs. As a growing number of attacks have been reported to generate adversarial inputs of varying sophistication, the defense-attack arms race has been accelerated. In this paper, we present MODEF, a cross-layer model diversity ensemble framework. MODEF intelligently combines unsupervised model denoising ensemble with supervised model verification ensemble by quantifying model diversity, aiming to boost the robustness of the target model against adversarial examples. Evaluated using eleven representative attacks on popular benchmark datasets, we show that MODEF achieves remarkable defense success rates, compared with existing defense methods, and provides a superior capability of repairing adversarial inputs and making correct predictions with high accuracy in the presence of black-box attacks.

Learning Rate (LR) is an important hyper-parameter to tune for effective training of deep neural networks (DNNs). Even for the baseline of a constant learning rate, it is non-trivial to choose a good constant value for training a DNN. Dynamic learning rates involve multi-step tuning of LR values at various stages of the training process and offer high accuracy and fast convergence. However, they are much harder to tune. In this paper, we present a comprehensive study of 13 learning rate functions and their associated LR policies by examining their range parameters, step parameters, and value update parameters. We propose a set of metrics for evaluating and selecting LR policies, including the classification confidence, variance, cost, and robustness, and implement them in LRBench, an LR benchmarking system. LRBench can assist end-users and DNN developers to select good LR policies and avoid bad LR policies for training their DNNs. We tested LRBench on Caffe, an open source deep learning framework, to showcase the tuning optimization of LR policies. Evaluated through extensive experiments, we attempt to demystify the tuning of LR policies by identifying good LR policies with effective LR value ranges and step sizes for LR update schedules.