Searcharxiv⌕ Search

arXiv subjects

David J. Miller

Publications and source records attributed to David J. Miller.

At least 19 recordsLinked to original sources
arXiv2026

CSO-LLM: Class Subspace Orthogonalization for Post-Training Backdoor Detection and Trigger Inversion in LLMs

While post-training backdoor detection and trigger inversion schemes have been developed for AIs used e.g. for images, there is a paucity of such methods for LLMs. First, the LLM input space is discrete, with up to 150,000^k k-tuples to consider with k the token-length of a putative trigger. Second, one must blacklist tokens typical of the putative target response (class) of an attack, as such tokens may give false detection signals. However, a comprehensive blacklist is not available, in general, for a given domain. We develop a highly effective detection and inversion framework for LLMs treated as classifiers. Central to our approach is class subspace orthogonalization (CSO), a novel plug-and-play paradigm for backdoor detection that serves two fundamental roles when applied to LLMs: i) it enhances both sensitivity and specificity of a baseline detector; ii) it provides a form of implicit blacklisting, as it penalizes against inclusion, in a candidate trigger, of tokens that induce signal perturbations "in the direction of" the putative target class of an attack. One version of our detector performs continuous optimization in token embedding space, while a companion trigger-inversion and detection method performs greedy accretion in discrete token space. Our methods give both strong detection performance and accurate inversion of ground-truth triggers on several LLM classification domains, and for several different LLM architectures.

cs.CR↗
arXiv2026

A Novel Latent-Class Attack and its Detection by Class Subspace Orthogonalization

Deep learning, which in general relies on voluminous amounts of training data, is vulnerable to data poisoning attacks, including error-generic attacks and backdoors (Trojans). In this work, we propose a new data poisoning attack we dub a latent class attack. Here, all poisoned examples are from a class that is novel (unknown) for the given classification domain and are mislabeled to one of the known classes (the target class) of the domain, so that the model learns to recognize the novel class as a sub-class of the target class. Such attacks could be used e.g. to defeat AI-based access control systems, or could cause a "foe" to be classified as a "friend". We also propose a post-training defense to detect this attack, without any access to the training set. This detection approach builds on "class subspace orthogonalization" (CSO), a plug-and-play paradigm demonstrated to improve existing backdoor detectors. Here, CSO is used to seek an input (a putative unknown class instance) whose internal representation is not aligned with any of the known classes, and yet which is classified with confidence to one of these classes. Finally, specific to image classification domains, we propose a method for visualizing the estimated unknown class instance, providing explainability to our latent class detections.

cs.LG↗
arXiv2025

Improving the Sensitivity of Backdoor Detectors via Class Subspace Orthogonalization

Most post-training backdoor detection methods rely on attacked models exhibiting extreme outlier detection statistics for the target class of an attack, compared to non-target classes. However, these approaches may fail: (1) when some (non-target) classes are easily discriminable from all others, in which case they may naturally achieve extreme detection statistics (e.g., decision confidence); and (2) when the backdoor is subtle, i.e., with its features weak relative to intrinsic class-discriminative features. A key observation is that the backdoor target class has contributions to its detection statistic from both the backdoor trigger and from its intrinsic features, whereas non-target classes only have contributions from their intrinsic features. To achieve more sensitive detectors, we thus propose to suppress intrinsic features while optimizing the detection statistic for a given class. For non-target classes, such suppression will drastically reduce the achievable statistic, whereas for the target class the (significant) contribution from the backdoor trigger remains. In practice, we formulate a constrained optimization problem, leveraging a small set of clean examples from a given class, and optimizing the detection statistic while orthogonalizing with respect to the class's intrinsic features. We dub this plug-and-play approach Class Subspace Orthogonalization (CSO) and assess it against challenging mixed-label and adaptive attacks.

cs.LG↗
arXiv2025

Inverting Trojans in LLMs

While effective backdoor detection and inversion schemes have been developed for AIs used e.g. for images, there are challenges in "porting" these methods to LLMs. First, the LLM input space is discrete, which precludes gradient-based search over this space, central to many backdoor inversion methods. Second, there are ~30,000^k k-tuples to consider, k the token-length of a putative trigger. Third, for LLMs there is the need to blacklist tokens that have strong marginal associations with the putative target response (class) of an attack, as such tokens give false detection signals. However, good blacklists may not exist for some domains. We propose a LLM trigger inversion approach with three key components: i) discrete search, with putative triggers greedily accreted, starting from a select list of singletons; ii) implicit blacklisting, achieved by evaluating the average cosine similarity, in activation space, between a candidate trigger and a small clean set of samples from the putative target class; iii) detection when a candidate trigger elicits high misclassifications, and with unusually high decision confidence. Unlike many recent works, we demonstrate that our approach reliably detects and successfully inverts ground-truth backdoor trigger phrases.

cs.LG↗
arXiv2025

CEPA: Consensus Embedded Perturbation for Agnostic Detection and Inversion of Backdoors

A variety of defenses have been proposed against Trojans planted in (backdoor attacks on) deep neural network (DNN) classifiers. Backdoor-agnostic methods seek to reliably detect and/or to mitigate backdoors irrespective of the incorporation mechanism used by the attacker, while inversion methods explicitly assume one. In this paper, we describe a new detector that: relies on embedded feature representations to estimate (invert) the backdoor and to identify its target class; can operate without access to the training dataset; and is highly effective for various incorporation mechanisms (i.e., is backdoor agnostic). Our detection approach is evaluated -- and found to be favorable - in comparison with an array of published defenses for a variety of different attacks on the CIFAR-10 and CIFAR-100 image-classification domains.

cs.CR↗
arXiv2024

A Study of Backdoors in Instruction Fine-tuned Language Models

Backdoor data poisoning, inserted within instruction examples used to fine-tune a foundation Large Language Model (LLM) for downstream tasks (\textit{e.g.,} sentiment prediction), is a serious security concern due to the evasive nature of such attacks. The poisoning is usually in the form of a (seemingly innocuous) trigger word or phrase inserted into a very small fraction of the fine-tuning samples from a target class. Such backdoor attacks can: alter response sentiment, violate censorship, over-refuse (invoke censorship for legitimate queries), inject false content, or trigger nonsense responses (hallucinations). In this work we investigate the efficacy of instruction fine-tuning backdoor attacks as attack "hyperparameters" are varied under a variety of scenarios, considering: the trigger location in the poisoned examples; robustness to change in the trigger location, partial triggers, and synonym substitutions at test time; attack transfer from one (fine-tuning) domain to a related test domain; and clean-label vs. dirty-label poisoning. Based on our observations, we propose and evaluate two defenses against these attacks: i) a \textit{during-fine-tuning defense} based on word-frequency counts that assumes the (possibly poisoned) fine-tuning dataset is available and identifies the backdoor trigger tokens; and ii) a \textit{post-fine-tuning defense} based on downstream clean fine-tuning of the backdoored LLM with a small defense dataset. Finally, we provide a brief survey of related work on backdoor attacks and defenses.

cs.CR↗
arXiv2023

Post-Training Overfitting Mitigation in DNN Classifiers

Well-known (non-malicious) sources of overfitting in deep neural net (DNN) classifiers include: i) large class imbalances; ii) insufficient training-set diversity; and iii) over-training. In recent work, it was shown that backdoor data-poisoning also induces overfitting, with unusually large classification margins to the attacker's target class, mediated particularly by (unbounded) ReLU activations that allow large signals to propagate in the DNN. Thus, an effective post-training (with no knowledge of the training set or training process) mitigation approach against backdoors was proposed, leveraging a small clean dataset, based on bounding neural activations. Improving upon that work, we threshold activations specifically to limit maximum margins (MMs), which yields performance gains in backdoor mitigation. We also provide some analytical support for this mitigation approach. Most importantly, we show that post-training MM-based regularization substantially mitigates non-malicious overfitting due to class imbalances and overtraining. Thus, unlike adversarial training, which provides some resilience against attacks but which harms clean (attack-free) generalization, we demonstrate an approach originating from adversarial learning that helps clean generalization accuracy. Experiments on CIFAR-10 and CIFAR-100, in comparison with peer methods, demonstrate strong performance of our methods.

cs.LG↗
arXiv2023

Backdoor Mitigation by Correcting the Distribution of Neural Activations

Backdoor (Trojan) attacks are an important type of adversarial exploit against deep neural networks (DNNs), wherein a test instance is (mis)classified to the attacker's target class whenever the attacker's backdoor trigger is present. In this paper, we reveal and analyze an important property of backdoor attacks: a successful attack causes an alteration in the distribution of internal layer activations for backdoor-trigger instances, compared to that for clean instances. Even more importantly, we find that instances with the backdoor trigger will be correctly classified to their original source classes if this distribution alteration is corrected. Based on our observations, we propose an efficient and effective method that achieves post-training backdoor mitigation by correcting the distribution alteration using reverse-engineered triggers. Notably, our method does not change any trainable parameters of the DNN, but achieves generally better mitigation performance than existing methods that do require intensive DNN parameter tuning. It also efficiently detects test instances with the trigger, which may help to catch adversarial entities in the act of exploiting the backdoor.

cs.LG↗
arXiv2023

Improved Activation Clipping for Universal Backdoor Mitigation and Test-Time Detection

Deep neural networks are vulnerable to backdoor attacks (Trojans), where an attacker poisons the training set with backdoor triggers so that the neural network learns to classify test-time triggers to the attacker's designated target class. Recent work shows that backdoor poisoning induces over-fitting (abnormally large activations) in the attacked model, which motivates a general, post-training clipping method for backdoor mitigation, i.e., with bounds on internal-layer activations learned using a small set of clean samples. We devise a new such approach, choosing the activation bounds to explicitly limit classification margins. This method gives superior performance against peer methods for CIFAR-10 image classification. We also show that this method has strong robustness against adaptive attacks, X2X attacks, and on different datasets. Finally, we demonstrate a method extension for test-time detection and correction based on the output differences between the original and activation-bounded networks. The code of our method is online available.

cs.LG↗
arXiv2023

MM-BD: Post-Training Detection of Backdoor Attacks with Arbitrary Backdoor Pattern Types Using a Maximum Margin Statistic

Backdoor attacks are an important type of adversarial threat against deep neural network classifiers, wherein test samples from one or more source classes will be (mis)classified to the attacker's target class when a backdoor pattern is embedded. In this paper, we focus on the post-training backdoor defense scenario commonly considered in the literature, where the defender aims to detect whether a trained classifier was backdoor-attacked without any access to the training set. Many post-training detectors are designed to detect attacks that use either one or a few specific backdoor embedding functions (e.g., patch-replacement or additive attacks). These detectors may fail when the backdoor embedding function used by the attacker (unknown to the defender) is different from the backdoor embedding function assumed by the defender. In contrast, we propose a post-training defense that detects backdoor attacks with arbitrary types of backdoor embeddings, without making any assumptions about the backdoor embedding type. Our detector leverages the influence of the backdoor attack, independent of the backdoor embedding mechanism, on the landscape of the classifier's outputs prior to the softmax layer. For each class, a maximum margin statistic is estimated. Detection inference is then performed by applying an unsupervised anomaly detector to these statistics. Thus, our detector does not need any legitimate clean samples, and can efficiently detect backdoor attacks with arbitrary numbers of source classes. These advantages over several state-of-the-art methods are demonstrated on four datasets, for three different types of backdoor patterns, and for a variety of attack configurations. Finally, we propose a novel, general approach for backdoor mitigation once a detection is made. The mitigation approach was the runner-up at the first IEEE Trojan Removal Competition. The code is online available.

cs.LG↗
arXiv2023

Spatial mapping and analysis of graphene nanomechanical resonator networks

Nanoelectromechanical (NEMS) resonator networks have drawn increasing interest due to their potential applications in emergent behavior, sensing, phononics, and mechanical information processing. A challenge toward realizing these large-scale networks is the ability to controllably tune and reconfigure the collective, macroscopic properties of the network, which relies directly on the development of methods to characterize the constituent NEMS resonator building blocks and their coupling. In this work, we demonstrate a scalable optical technique to spatially map graphene NEMS networks and read out the fixed-frequency collective response as a single vector. Using the response vectors, we introduce an efficient algebraic approach to quantify the site-specific elasticity, mass, damping, and coupling parameters of network clusters. We apply this technique to accurately characterize single uncoupled resonators and coupled resonator pairs by sampling them at just two frequencies, and without the use of curve fitting or the associated a priori parameter estimates. Our technique may be applied to a range of classical and quantum resonator systems and fills in a vital gap for programming NEMS networks.

cond-mat.mes-hall↗
arXiv2022

Anomaly Detection of Adversarial Examples using Class-conditional Generative Adversarial Networks

Deep Neural Networks (DNNs) have been shown vulnerable to Test-Time Evasion attacks (TTEs, or adversarial examples), which, by making small changes to the input, alter the DNN's decision. We propose an unsupervised attack detector on DNN classifiers based on class-conditional Generative Adversarial Networks (GANs). We model the distribution of clean data conditioned on the predicted class label by an Auxiliary Classifier GAN (AC-GAN). Given a test sample and its predicted class, three detection statistics are calculated based on the AC-GAN Generator and Discriminator. Experiments on image classification datasets under various TTE attacks show that our method outperforms previous detection methods. We also investigate the effectiveness of anomaly detection using different DNN layers (input features or internal-layer features) and demonstrate, as one might expect, that anomalies are harder to detect using features closer to the DNN's output layer.

cs.LG↗
arXiv2022

A BIC-based Mixture Model Defense against Data Poisoning Attacks on Classifiers

Data Poisoning (DP) is an effective attack that causes trained classifiers to misclassify their inputs. DP attacks significantly degrade a classifier's accuracy by covertly injecting attack samples into the training set. Broadly applicable to different classifier structures, without strong assumptions about the attacker, an {\it unsupervised} Bayesian Information Criterion (BIC)-based mixture model defense against "error generic" DP attacks is herein proposed that: 1) addresses the most challenging {\it embedded} DP scenario wherein, if DP is present, the poisoned samples are an {\it a priori} unknown subset of the training set, and with no clean validation set available; 2) applies a mixture model both to well-fit potentially multi-modal class distributions and to capture poisoned samples within a small subset of the mixture components; 3) jointly identifies poisoned components and samples by minimizing the BIC cost defined over the whole training set, with the identified poisoned data removed prior to classifier training. Our experimental results, for various classifier structures and benchmark datasets, demonstrate the effectiveness and universality of our defense under strong DP attacks, as well as its superiority over other works.

cs.LG↗
arXiv2022

Improved Constraints on Effective Top Quark Interactions using Edge Convolution Networks

We explore the potential of Graph Neural Networks (GNNs) to improve the performance of high-dimensional effective field theory parameter fits to collider data beyond traditional rectangular cut-based differential distribution analyses. In this study, we focus on a SMEFT analysis of $pp \to t\bar t$ production, including top decays, where the linear effective field deformation is parametrised by thirteen independent Wilson coefficients. The application of GNNs allows us to condense the multidimensional phase space information available for the discrimination of BSM effects from the SM expectation by considering all available final state correlations directly. The number of contributing new physics couplings very quickly leads to statistical limitations when the GNN output is directly employed as an EFT discrimination tool. However, a selection based on minimising the SM contribution enhances the fit's sensitivity when reflected as a (non-rectangular) selection on the inclusive data samples that are typically employed when looking for non-resonant deviations from the SM by means of differential distributions.

hep-ph↗
arXiv2022

Post-Training Detection of Backdoor Attacks for Two-Class and Multi-Attack Scenarios

Backdoor attacks (BAs) are an emerging threat to deep neural network classifiers. A victim classifier will predict to an attacker-desired target class whenever a test sample is embedded with the same backdoor pattern (BP) that was used to poison the classifier's training set. Detecting whether a classifier is backdoor attacked is not easy in practice, especially when the defender is, e.g., a downstream user without access to the classifier's training set. This challenge is addressed here by a reverse-engineering defense (RED), which has been shown to yield state-of-the-art performance in several domains. However, existing REDs are not applicable when there are only {\it two classes} or when {\it multiple attacks} are present. These scenarios are first studied in the current paper, under the practical constraints that the defender neither has access to the classifier's training set nor to supervision from clean reference classifiers trained for the same domain. We propose a detection framework based on BP reverse-engineering and a novel {\it expected transferability} (ET) statistic. We show that our ET statistic is effective {\it using the same detection threshold}, irrespective of the classification domain, the attack configuration, and the BP reverse-engineering algorithm that is used. The excellent performance of our method is demonstrated on six benchmark datasets. Notably, our detection framework is also applicable to multi-class scenarios with multiple attacks. Code is available at https://github.com/zhenxianglance/2ClassBADetection.

cs.CR↗
arXiv2021

Test-Time Detection of Backdoor Triggers for Poisoned Deep Neural Networks

Backdoor (Trojan) attacks are emerging threats against deep neural networks (DNN). A DNN being attacked will predict to an attacker-desired target class whenever a test sample from any source class is embedded with a backdoor pattern; while correctly classifying clean (attack-free) test samples. Existing backdoor defenses have shown success in detecting whether a DNN is attacked and in reverse-engineering the backdoor pattern in a "post-training" regime: the defender has access to the DNN to be inspected and a small, clean dataset collected independently, but has no access to the (possibly poisoned) training set of the DNN. However, these defenses neither catch culprits in the act of triggering the backdoor mapping, nor mitigate the backdoor attack at test-time. In this paper, we propose an "in-flight" defense against backdoor attacks on image classification that 1) detects use of a backdoor trigger at test-time; and 2) infers the class of origin (source class) for a detected trigger example. The effectiveness of our defense is demonstrated experimentally against different strong backdoor attacks.

cs.CR↗
arXiv2021

Detecting Backdoor Attacks Against Point Cloud Classifiers

Backdoor attacks (BA) are an emerging threat to deep neural network classifiers. A classifier being attacked will predict to the attacker's target class when a test sample from a source class is embedded with the backdoor pattern (BP). Recently, the first BA against point cloud (PC) classifiers was proposed, creating new threats to many important applications including autonomous driving. Such PC BAs are not detectable by existing BA defenses due to their special BP embedding mechanism. In this paper, we propose a reverse-engineering defense that infers whether a PC classifier is backdoor attacked, without access to its training set or to any clean classifiers for reference. The effectiveness of our defense is demonstrated on the benchmark ModeNet40 dataset for PCs.

cs.CR↗
arXiv2021

Backdoor Attack and Defense for Deep Regression

We demonstrate a backdoor attack on a deep neural network used for regression. The backdoor attack is localized based on training-set data poisoning wherein the mislabeled samples are surrounded by correctly labeled ones. We demonstrate how such localization is necessary for attack success. We also study the performance of a backdoor defense using gradient-based discovery of local error maximizers. Local error maximizers which are associated with significant (interpolation) error, and are proximal to many training samples, are suspicious. This method is also used to accurately train for deep regression in the first place by active (deep) learning leveraging an "oracle" capable of providing real-valued supervision (a regression target) for samples. Such oracles, including traditional numerical solvers of PDEs or SDEs using finite difference or Monte Carlo approximations, are far more computationally costly compared to deep regression.

cs.LG↗