CERT-ED: Certifiably Robust Text Classification for Edit Distance

With the growing integration of AI in daily life, ensuring the robustness of systems to inference-time attacks is crucial. Among the approaches for certifying robustness to such adversarial examples, randomized smoothing has emerged as highly promising due to its nature as a wrapper around arbitrary black-box models. Previous work on randomized smoothing in natural language processing has primarily focused on specific subsets of edit distance operations, such as synonym substitution or word insertion, without exploring the certification of all edit operations. In this paper, we adapt Randomized Deletion (Huang et al., 2023) and propose, CERTified Edit Distance defense (CERT-ED) for natural language classification. Through comprehensive experiments, we demonstrate that CERT-ED outperforms the existing Hamming distance method RanMASK (Zeng et al., 2023) in 4 out of 5 datasets in terms of both accuracy and the cardinality of the certificate. By covering various threat models, including 5 direct and 5 transfer attacks, our method improves empirical robustness in 38 out of 50 settings.

Adaptive Data Analysis for Growing Data

Reuse of data in adaptive workflows poses challenges regarding overfitting and the statistical validity of results. Previous work has demonstrated that interacting with data via differentially private algorithms can mitigate overfitting, achieving worst-case generalization guarantees with asymptotically optimal data requirements. However, such past work assumes data is static and cannot accommodate situations where data grows over time. In this paper we address this gap, presenting the first generalization bounds for adaptive analysis in the dynamic data setting. We allow the analyst to adaptively schedule their queries conditioned on the current size of the data, in addition to previous queries and responses. We also incorporate time-varying empirical accuracy bounds and mechanisms, allowing for tighter guarantees as data accumulates. In a batched query setting, the asymptotic data requirements of our bound grows with the square-root of the number of adaptive queries, matching prior works' improvement over data splitting for the static setting. We instantiate our bound for statistical queries with the clipped Gaussian mechanism, where it empirically outperforms baselines composed from static bounds.

SEEP: Training Dynamics Grounds Latent Representation Search for Mitigating Backdoor Poisoning Attacks

Modern NLP models are often trained on public datasets drawn from diverse sources, rendering them vulnerable to data poisoning attacks. These attacks can manipulate the model's behavior in ways engineered by the attacker. One such tactic involves the implantation of backdoors, achieved by poisoning specific training instances with a textual trigger and a target class label. Several strategies have been proposed to mitigate the risks associated with backdoor attacks by identifying and removing suspected poisoned examples. However, we observe that these strategies fail to offer effective protection against several advanced backdoor attacks. To remedy this deficiency, we propose a novel defensive mechanism that first exploits training dynamics to identify poisoned samples with high precision, followed by a label propagation step to improve recall and thus remove the majority of poisoned instances. Compared with recent advanced defense methods, our method considerably reduces the success rates of several backdoor attacks while maintaining high classification accuracy on clean test sets.

RS-Reg: Probabilistic and Robust Certified Regression Through Randomized Smoothing

Randomized smoothing has shown promising certified robustness against adversaries in classification tasks. Despite such success with only zeroth-order access to base models, randomized smoothing has not been extended to a general form of regression. By defining robustness in regression tasks flexibly through probabilities, we demonstrate how to establish upper bounds on input data point perturbation (using the $\ell_2$ norm) for a user-specified probability of observing valid outputs. Furthermore, we showcase the asymptotic property of a basic averaging function in scenarios where the regression model operates without any constraint. We then derive a certified upper bound of the input perturbations when dealing with a family of regression models where the outputs are bounded. Our simulations verify the validity of the theoretical results and reveal the advantages and limitations of simple smoothing functions, i.e., averaging, in regression tasks. The code is publicly available at \url{https://github.com/arekavandi/Certified_Robust_Regression}.

Transferring Troubles: Cross-Lingual Transferability of Backdoor Attacks in LLMs with Instruction Tuning

The implications of backdoor attacks on English-centric large language models (LLMs) have been widely examined - such attacks can be achieved by embedding malicious behaviors during training and activated under specific conditions that trigger malicious outputs. However, the impact of backdoor attacks on multilingual models remains under-explored. Our research focuses on cross-lingual backdoor attacks against multilingual LLMs, particularly investigating how poisoning the instruction-tuning data in one or two languages can affect the outputs in languages whose instruction-tuning data was not poisoned. Despite its simplicity, our empirical analysis reveals that our method exhibits remarkable efficacy in models like mT5, BLOOM, and GPT-3.5-turbo, with high attack success rates, surpassing 95% in several languages across various scenarios. Alarmingly, our findings also indicate that larger models show increased susceptibility to transferable cross-lingual backdoor attacks, which also applies to LLMs predominantly pre-trained on English data, such as Llama2, Llama3, and Gemma. Moreover, our experiments show that triggers can still work even after paraphrasing, and the backdoor mechanism proves highly effective in cross-lingual response settings across 25 languages, achieving an average attack success rate of 50%. Our study aims to highlight the vulnerabilities and significant security risks present in current multilingual LLMs, underscoring the emergent need for targeted security measures.

Backdoor Attack on Multilingual Machine Translation

While multilingual machine translation (MNMT) systems hold substantial promise, they also have security vulnerabilities. Our research highlights that MNMT systems can be susceptible to a particularly devious style of backdoor attack, whereby an attacker injects poisoned data into a low-resource language pair to cause malicious translations in other languages, including high-resource languages. Our experimental results reveal that injecting less than 0.01% poisoned data into a low-resource language pair can achieve an average 20% attack success rate in attacking high-resource language pairs. This type of attack is of particular concern, given the larger attack surface of languages inherent to low-resource settings. Our aim is to bring attention to these vulnerabilities within MNMT systems with the hope of encouraging the community to address security concerns in machine translation, especially in the context of low-resource languages.

It's Simplex! Disaggregating Measures to Improve Certified Robustness

Certified robustness circumvents the fragility of defences against adversarial attacks, by endowing model predictions with guarantees of class invariance for attacks up to a calculated size. While there is value in these certifications, the techniques through which we assess their performance do not present a proper accounting of their strengths and weaknesses, as their analysis has eschewed consideration of performance over individual samples in favour of aggregated measures. By considering the potential output space of certified models, this work presents two distinct approaches to improve the analysis of certification mechanisms, that allow for both dataset-independent and dataset-dependent measures of certification performance. Embracing such a perspective uncovers new certification approaches, which have the potential to more than double the achievable radius of certification, relative to current state-of-the-art. Empirical evaluation verifies that our new approach can certify $9\%$ more samples at noise scale $\sigma = 1$, with greater relative improvements observed as the difficulty of the predictive task increases.

Enhancing the Antidote: Improved Pointwise Certifications against Poisoning Attacks

Poisoning attacks can disproportionately influence model behaviour by making small changes to the training corpus. While defences against specific poisoning attacks do exist, they in general do not provide any guarantees, leaving them potentially countered by novel attacks. In contrast, by examining worst-case behaviours Certified Defences make it possible to provide guarantees of the robustness of a sample against adversarial attacks modifying a finite number of training samples, known as pointwise certification. We achieve this by exploiting both Differential Privacy and the Sampled Gaussian Mechanism to ensure the invariance of prediction for each testing instance against finite numbers of poisoned examples. In doing so, our model provides guarantees of adversarial robustness that are more than twice as large as those provided by prior certifications.

Exploiting Certified Defences to Attack Randomised Smoothing

In guaranteeing that no adversarial examples exist within a bounded region, certification mechanisms play an important role in neural network robustness. Concerningly, this work demonstrates that the certification mechanisms themselves introduce a new, heretofore undiscovered attack surface, that can be exploited by attackers to construct smaller adversarial perturbations. While these attacks exist outside the certification region in no way invalidate certifications, minimising a perturbation's norm significantly increases the level of difficulty associated with attack detection. In comparison to baseline attacks, our new framework yields smaller perturbations more than twice as frequently as any other approach, resulting in an up to $34 \%$ reduction in the median perturbation norm. That this approach also requires $90 \%$ less computational time than approaches like PGD. That these reductions are possible suggests that exploiting this new attack vector would allow attackers to more frequently construct hard to detect adversarial attacks, by exploiting the very systems designed to defend deployed models.

Certified Robustness of Learning-based Static Malware Detectors

Certified defenses are a recent development in adversarial machine learning (ML), which aim to rigorously guarantee the robustness of ML models to adversarial perturbations. A large body of work studies certified defenses in computer vision, where $\ell_p$ norm-bounded evasion attacks are adopted as a tractable threat model. However, this threat model has known limitations in vision, and is not applicable to other domains -- e.g., where inputs may be discrete or subject to complex constraints. Motivated by this gap, we study certified defenses for malware detection, a domain where attacks against ML-based systems are a real and current threat. We consider static malware detection systems that operate on byte-level data. Our certified defense is based on the approach of randomized smoothing which we adapt by: (1) replacing the standard Gaussian randomization scheme with a novel deletion randomization scheme that operates on bytes or chunks of an executable; and (2) deriving a certificate that measures robustness to evasion attacks in terms of generalized edit distance. To assess the size of robustness certificates that are achievable while maintaining high accuracy, we conduct experiments on malware datasets using a popular convolutional malware detection model, MalConv. We are able to accurately classify 91% of the inputs while being certifiably robust to any adversarial perturbations of edit distance 128 bytes or less. By comparison, an existing certification of up to 128 bytes of substitutions (without insertions or deletions) achieves an accuracy of 78%. In addition, given that robustness certificates are conservative, we evaluate practical robustness to several recently published evasion attacks and, in some cases, find robustness beyond certified guarantees.