Recent work in open-domain question answering (ODQA) has shown that adversarial poisoning of the search collection can cause large drops in accuracy for production systems. However, little to no work has proposed methods to defend against these attacks. To do so, we rely on the intuition that redundant information often exists in large corpora. To find it, we introduce a method that uses query augmentation to search for a diverse set of passages that could answer the original question but are less likely to have been poisoned. We integrate these new passages into the model through the design of a novel confidence method, comparing the predicted answer to its appearance in the retrieved contexts (what we call \textit{Confidence from Answer Redundancy}, i.e. CAR). Together these methods allow for a simple but effective way to defend against poisoning attacks that provides gains of nearly 20\% exact match across varying levels of data poisoning/knowledge conflicts.
相關內容
ML models are known to be vulnerable to adversarial query attacks. In these attacks, queries are iteratively perturbed towards a particular class without any knowledge of the target model besides its output. The prevalence of remotely-hosted ML classification models and Machine-Learning-as-a-Service platforms means that query attacks pose a real threat to the security of these systems. To deal with this, stateful defenses have been proposed to detect query attacks and prevent the generation of adversarial examples by monitoring and analyzing the sequence of queries received by the system. Several stateful defenses have been proposed in recent years. However, these defenses rely solely on similarity or out-of-distribution detection methods that may be effective in other domains. In the malware detection domain, the methods to generate adversarial examples are inherently different, and therefore we find that such detection mechanisms are significantly less effective. Hence, in this paper, we present MalProtect, which is a stateful defense against query attacks in the malware detection domain. MalProtect uses several threat indicators to detect attacks. Our results show that it reduces the evasion rate of adversarial query attacks by 80+\% in Android and Windows malware, across a range of attacker scenarios. In the first evaluation of its kind, we show that MalProtect outperforms prior stateful defenses, especially under the peak adversarial threat.
Several companies often safeguard their trained deep models (i.e., details of architecture, learnt weights, training details etc.) from third-party users by exposing them only as black boxes through APIs. Moreover, they may not even provide access to the training data due to proprietary reasons or sensitivity concerns. In this work, we propose a novel defense mechanism for black box models against adversarial attacks in a data-free set up. We construct synthetic data via generative model and train surrogate network using model stealing techniques. To minimize adversarial contamination on perturbed samples, we propose 'wavelet noise remover' (WNR) that performs discrete wavelet decomposition on input images and carefully select only a few important coefficients determined by our 'wavelet coefficient selection module' (WCSM). To recover the high-frequency content of the image after noise removal via WNR, we further train a 'regenerator' network with the objective of retrieving the coefficients such that the reconstructed image yields similar to original predictions on the surrogate model. At test time, WNR combined with trained regenerator network is prepended to the black box network, resulting in a high boost in adversarial accuracy. Our method improves the adversarial accuracy on CIFAR-10 by 38.98% and 32.01% on state-of-the-art Auto Attack compared to baseline, even when the attacker uses surrogate architecture (Alexnet-half and Alexnet) similar to the black box architecture (Alexnet) with same model stealing strategy as defender. The code is available at //github.com/vcl-iisc/data-free-black-box-defense
Uses of artificial intelligence (AI), especially those powered by machine learning approaches, are growing in sectors and societies around the world. How will AI adoption proceed, especially in the international security realm? Research on automation bias suggests that humans can often be overconfident in AI, whereas research on algorithm aversion shows that, as the stakes of a decision rise, humans become more cautious about trusting algorithms. We theorize about the relationship between background knowledge about AI, trust in AI, and how these interact with other factors to influence the probability of automation bias in the international security context. We test these in a preregistered task identification experiment across a representative sample of 9000 adults in 9 countries with varying levels of AI industries. The results strongly support the theory, especially concerning AI background knowledge. A version of the Dunning Kruger effect appears to be at play, whereby those with the lowest level of experience with AI are slightly more likely to be algorithm-averse, then automation bias occurs at lower levels of knowledge before leveling off as a respondent's AI background reaches the highest levels. Additional results show effects from the task's difficulty, overall AI trust, and whether a human or AI decision aid is described as highly competent or less competent.
The increasing access to data poses both opportunities and risks in deep learning, as one can manipulate the behaviors of deep learning models with malicious training samples. Such attacks are known as data poisoning. Recent advances in defense strategies against data poisoning have highlighted the effectiveness of aggregation schemes in achieving state-of-the-art results in certified poisoning robustness. However, the practical implications of these approaches remain unclear. Here we focus on Deep Partition Aggregation, a representative aggregation defense, and assess its practical aspects, including efficiency, performance, and robustness. For evaluations, we use ImageNet resized to a resolution of 64 by 64 to enable evaluations at a larger scale than previous ones. Firstly, we demonstrate a simple yet practical approach to scaling base models, which improves the efficiency of training and inference for aggregation defenses. Secondly, we provide empirical evidence supporting the data-to-complexity ratio, i.e. the ratio between the data set size and sample complexity, as a practical estimation of the maximum number of base models that can be deployed while preserving accuracy. Last but not least, we point out how aggregation defenses boost poisoning robustness empirically through the poisoning overfitting phenomenon, which is the key underlying mechanism for the empirical poisoning robustness of aggregations. Overall, our findings provide valuable insights for practical implementations of aggregation defenses to mitigate the threat of data poisoning.
Data poisoning considers cases when an adversary manipulates the behavior of machine learning algorithms through malicious training data. Existing threat models of data poisoning center around a single metric, the number of poisoned samples. In consequence, if attackers can poison more samples than expected with affordable overhead, as in many practical scenarios, they may be able to render existing defenses ineffective in a short time. To address this issue, we leverage timestamps denoting the birth dates of data, which are often available but neglected in the past. Benefiting from these timestamps, we propose a temporal threat model of data poisoning with two novel metrics, earliness and duration, which respectively measure how long an attack started in advance and how long an attack lasted. Using these metrics, we define the notions of temporal robustness against data poisoning, providing a meaningful sense of protection even with unbounded amounts of poisoned samples. We present a benchmark with an evaluation protocol simulating continuous data collection and periodic deployments of updated models, thus enabling empirical evaluation of temporal robustness. Lastly, we develop and also empirically verify a baseline defense, namely temporal aggregation, offering provable temporal robustness and highlighting the potential of our temporal threat model for data poisoning.
We introduce the Condorcet attack, a new threat to fair transaction ordering. Specifically, the attack undermines batch-order-fairness, the strongest notion of transaction fair ordering proposed to date. The batch-order-fairness guarantees that a transaction tx is ordered before tx' if a majority of nodes in the system receive tx before tx'; the only exception (due to an impossibility result) is when tx and tx' fall into a so-called "Condorcet cycle". When this happens, tx and tx' along with other transactions within the cycle are placed in a batch, and any unfairness inside a batch is ignored. In the Condorcet attack, an adversary attempts to undermine the system's fairness by imposing Condorcet cycles to the system. In this work, we show that the adversary can indeed impose a Condorcet cycle by submitting as few as two otherwise legitimate transactions to the system. Remarkably, the adversary (e.g., a malicious client) can achieve this even when all the nodes in the system behave honestly. A notable feature of the attack is that it is capable of "trapping" transactions that do not naturally fall inside a cycle, i.e. those that are transmitted at significantly different times (with respect to the network latency). To mitigate the attack, we propose three methods based on three different complementary approaches. We show the effectiveness of the proposed mitigation methods through simulations, and explain their limitations.
Knowledge enhanced pre-trained language models (K-PLMs) are shown to be effective for many public tasks in the literature but few of them have been successfully applied in practice. To address this problem, we propose K-AID, a systematic approach that includes a low-cost knowledge acquisition process for acquiring domain knowledge, an effective knowledge infusion module for improving model performance, and a knowledge distillation component for reducing the model size and deploying K-PLMs on resource-restricted devices (e.g., CPU) for real-world application. Importantly, instead of capturing entity knowledge like the majority of existing K-PLMs, our approach captures relational knowledge, which contributes to better-improving sentence-level text classification and text matching tasks that play a key role in question answering (QA). We conducted a set of experiments on five text classification tasks and three text matching tasks from three domains, namely E-commerce, Government, and Film&TV, and performed online A/B tests in E-commerce. Experimental results show that our approach is able to achieve substantial improvement on sentence-level question answering tasks and bring beneficial business value in industrial settings.
Rehearsal, seeking to remind the model by storing old knowledge in lifelong learning, is one of the most effective ways to mitigate catastrophic forgetting, i.e., biased forgetting of previous knowledge when moving to new tasks. However, the old tasks of the most previous rehearsal-based methods suffer from the unpredictable domain shift when training the new task. This is because these methods always ignore two significant factors. First, the Data Imbalance between the new task and old tasks that makes the domain of old tasks prone to shift. Second, the Task Isolation among all tasks will make the domain shift toward unpredictable directions; To address the unpredictable domain shift, in this paper, we propose Multi-Domain Multi-Task (MDMT) rehearsal to train the old tasks and new task parallelly and equally to break the isolation among tasks. Specifically, a two-level angular margin loss is proposed to encourage the intra-class/task compactness and inter-class/task discrepancy, which keeps the model from domain chaos. In addition, to further address domain shift of the old tasks, we propose an optional episodic distillation loss on the memory to anchor the knowledge for each old task. Experiments on benchmark datasets validate the proposed approach can effectively mitigate the unpredictable domain shift.
As data are increasingly being stored in different silos and societies becoming more aware of data privacy issues, the traditional centralized training of artificial intelligence (AI) models is facing efficiency and privacy challenges. Recently, federated learning (FL) has emerged as an alternative solution and continue to thrive in this new reality. Existing FL protocol design has been shown to be vulnerable to adversaries within or outside of the system, compromising data privacy and system robustness. Besides training powerful global models, it is of paramount importance to design FL systems that have privacy guarantees and are resistant to different types of adversaries. In this paper, we conduct the first comprehensive survey on this topic. Through a concise introduction to the concept of FL, and a unique taxonomy covering: 1) threat models; 2) poisoning attacks and defenses against robustness; 3) inference attacks and defenses against privacy, we provide an accessible review of this important topic. We highlight the intuitions, key techniques as well as fundamental assumptions adopted by various attacks and defenses. Finally, we discuss promising future research directions towards robust and privacy-preserving federated learning.
Deep neural networks (DNN) have achieved unprecedented success in numerous machine learning tasks in various domains. However, the existence of adversarial examples has raised concerns about applying deep learning to safety-critical applications. As a result, we have witnessed increasing interests in studying attack and defense mechanisms for DNN models on different data types, such as images, graphs and text. Thus, it is necessary to provide a systematic and comprehensive overview of the main threats of attacks and the success of corresponding countermeasures. In this survey, we review the state of the art algorithms for generating adversarial examples and the countermeasures against adversarial examples, for the three popular data types, i.e., images, graphs and text.
小貼士
登錄享
相關主題
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191