In the rapidly growing digital economy, protecting intellectual property (IP) associated with digital products has become increasingly important. Within this context, machine learning (ML) models, being highly valuable digital assets, have gained significant attention for IP protection. This paper introduces a practical encryption-based framework called EncryIP, which seamlessly integrates a public-key encryption scheme into the model learning process. This approach enables the protected model to generate randomized and confused labels, ensuring that only individuals with accurate secret keys, signifying authorized users, can decrypt and reveal authentic labels. Importantly, the proposed framework not only facilitates the protected model to multiple authorized users without requiring repetitive training of the original ML model with IP protection methods but also maintains the model's performance without compromising its accuracy. Compared to existing methods like watermark-based, trigger-based, and passport-based approaches, EncryIP demonstrates superior effectiveness in both training protected models and efficiently detecting the unauthorized spread of ML models.
论文检索
输入标题、作者或关键词，从 11,272 篇学术成果中精准定位
Reinforcement Learning (RL) has achieved remarkable success in safety-critical areas, but it can be weakened by adversarial attacks. Recent studies have introduced ``smoothed policies" to enhance its robustness. Yet, it is still challenging to establish a provable guarantee to certify the bound of its total reward. Prior methods relied primarily on computing bounds using Lipschitz continuity or calculating the probability of cumulative reward being above specific thresholds. However, these techniques are only suited for continuous perturbations on the RL agent's observations and are restricted to perturbations bounded by the l2-norm. To address these limitations, this paper proposes a general black-box certification method, called ReCePS, which is capable of directly certifying the cumulative reward of the smoothed policy under various lp-norm bounded perturbations. Furthermore, we extend our methodology to certify perturbations on action spaces. Our approach leverages f-divergence to measure the distinction between the original distribution and the perturbed distribution, subsequently determining the certification bound by solving a convex optimisation problem. We provide a comprehensive theoretical analysis and run experiments in multiple environments. Our results show that our method not only improves the tightness of certified lower bound of the mean cumulative reward but also demonstrates better efficiency than state-of-the-art methods.
We tackle the problem of feature unlearning from a pre-trained image generative model: GANs and VAEs. Unlike a common unlearning task where an unlearning target is a subset of the training set, we aim to unlearn a specific feature, such as hairstyle from facial images, from the pre-trained generative models. As the target feature is only presented in a local region of an image, unlearning the entire image from the pre-trained model may result in losing other details in the remaining region of the image. To specify which features to unlearn, we collect randomly generated images that contain the target features. We then identify a latent representation corresponding to the target feature and then use the representation to fine-tune the pre-trained model. Through experiments on MNIST, CelebA, and FFHQ datasets, we show that target features are successfully removed while keeping the fidelity of the original models. Further experiments with an adversarial attack show that the unlearned model is more robust under the presence of malicious parties.
Safeguarded Progress in Reinforcement Learning: Safe Bayesian Exploration for Control Policy Synthesis
PDF ↗This paper addresses the problem of maintaining safety during training in Reinforcement Learning (RL), such that the safety constraint violations are bounded at any point during learning. As enforcing safety during training might severely limit the agent’s exploration, we propose here a new architecture that handles the trade-off between efficient progress and safety during exploration. As the exploration progresses, we update via Bayesian inference Dirichlet-Categorical models of the transition probabilities of the Markov decision process that describes the environment dynamics. We then propose a way to approximate moments of belief about the risk associated to the action selection policy. We demonstrate that this approach can be easily interleaved with RL and we present experimental results to showcase the performance of the overall architecture.
SentinelLMs: Encrypted Input Adaptation and Fine-Tuning of Language Models for Private and Secure Inference
PDF ↗This paper addresses the privacy and security concerns associated with deep neural language models, which serve as crucial components in various modern AI-based applications. These models are often used after being pre-trained and fine-tuned for specific tasks, with deployment on servers accessed through the internet. However, this introduces two fundamental risks: (a) the transmission of user inputs to the server via the network gives rise to interception vulnerabilities, and (b) privacy concerns emerge as organizations that deploy such models store user data with restricted context. To address this, we propose a novel method to adapt and fine-tune transformer-based language models on passkey-encrypted user-specific text. The original pre-trained language model first undergoes a quick adaptation (without any further pre-training) with a series of irreversible transformations applied to the tokenizer and token embeddings. This enables the model to perform inference on encrypted inputs while preventing reverse engineering of text from model parameters and intermediate outputs. After adaptation, models are fine-tuned on encrypted versions of existing training datasets. Experimental evaluation employing adapted versions of renowned models (e.g., BERT, RoBERTa) across established benchmark English and multilingual datasets for text classification and sequence labeling shows that encrypted models achieve performance parity with their original counterparts. This serves to safeguard performance, privacy, and security cohesively.
Artificial Intelligence (AI) models have become an integral part of modern society, significantly improving human lives. However, ensuring the reliability and safety of these models is of paramount importance. One critical aspect is the continuous monitoring and verification of model performance to prevent any potential risks. Real-time online evaluation of AI models is necessary to maintain their effectiveness and mitigate any harm caused by performance degradation. The traditional approach to model evaluation involves supervised methods that rely on manual labeling to compare results with model predictions. Unfortunately, this method is not suitable for online model monitoring due to its inherent lag and high cost. While there have been attempts to explore free-label model evaluation, these approaches often consider only the global features of the entire dataset. Additionally, they can only perform model evaluation based on a single dimension of model confidence or features. In this paper, we propose a novel approach called Divide-and-Aggregate Learning (DAL) for unsupervised model evaluation. Our method addresses the limitations of previous approaches by dividing the output of the model into buckets, capturing local information of the distribution. We then aggregate this local information to obtain global information and further represent the relationship between the distribution and model performance. Importantly, our method can simultaneously handle the confidence distribution and feature distribution of the model output. Extensive experiments have been conducted to demonstrate the effectiveness of our DAL model. The results show that our approach outperforms previous methods on four widely used datasets. We will make our source code publicly available.
Identifying safe areas is a key point to guarantee trust for systems that are based on Deep Neural Networks (DNNs). To this end, we introduce the AllDNN-Verification problem: given a safety property and a DNN, enumerate the set of all the regions of the property input domain which are safe, i.e., where the property does hold. Due to the #P-hardness of the problem, we propose an efficient approximation method called ε-ProVe. Our approach exploits a controllable underestimation of the output reachable sets obtained via statistical prediction of tolerance limits, and can provide a tight —with provable probabilistic guarantees— lower estimate of the safe areas. Our empirical evaluation on different standard benchmarks shows the scalability and effectiveness of our method, offering valuable insights for this new type of verification of DNNs.
Combining Graph Transformers Based Multi-Label Active Learning and Informative Data Augmentation for Chest Xray Classification
PDF ↗Informative sample selection in active learning (AL) helps a machine learning system attain optimum performance with minimum labeled samples, thus improving human-in-the-loop computer-aided diagnosis systems with limited labeled data. Data augmentation is highly effective for enlarging datasets with less labeled data. Combining informative sample selection and data augmentation should leverage their respective advantages and improve performance of AL systems. We propose a novel approach to combine informative sample selection and data augmentation for multi-label active learning. Conventional informative sample selection approaches have mostly focused on the single-label case which do not perform optimally in the multi-label setting. We improve upon state-of-the-art multi-label active learning techniques by representing disease labels as graph nodes, use graph attention transformers (GAT) to learn more effective inter-label relationships and identify most informative samples. We generate transformations of these informative samples which are also informative. Experiments on public chest xray datasets show improved results over state-of-the-art multi-label AL techniques in terms of classification performance, learning rates, and robustness. We also perform qualitative analysis to determine the realism of generated images.
Handling Long and Richly Constrained Tasks through Constrained Hierarchical Reinforcement Learning
PDF ↗Safety in goal directed Reinforcement Learning (RL) settings has typically been handled through constraints over trajectories and have demonstrated good performance in primarily short horizon tasks. In this paper, we are specifically interested in the problem of solving temporally extended decision making problems such as robots cleaning different areas in a house while avoiding slippery and unsafe areas (e.g., stairs) and retaining enough charge to move to a charging dock; in the presence of complex safety constraints. Our key contribution is a (safety) Constrained Search with Hierarchical Reinforcement Learning (CoSHRL) mechanism that combines an upper level constrained search agent (which computes a reward maximizing policy from a given start to a far away goal state while satisfying cost constraints) with a low-level goal conditioned RL agent (which estimates cost and reward values to move between nearby states). A major advantage of CoSHRL is that it can handle constraints on the cost value distribution (e.g., on Conditional Value at Risk, CVaR) and can adjust to flexible constraint thresholds without retraining. We perform extensive experiments with different types of safety constraints to demonstrate the utility of our approach over leading approaches in constrained and hierarchical RL.
Backdoors on federated learning will be diluted by subsequent benign updates. This is reflected in the significant reduction of attack success rate as iterations increase, ultimately failing. We use a new metric to quantify the degree of this weakened backdoor effect, called attack persistence. Given that research to improve this performance has not been widely noted, we propose a Full Combination Backdoor Attack (FCBA) method. It aggregates more combined trigger information for a more complete backdoor pattern in the global model. Trained backdoored global model is more resilient to benign updates, leading to a higher attack success rate on the test set. We test on three datasets and evaluate with two models across various settings. FCBA's persistence outperforms SOTA federated learning backdoor attacks. On GTSRB, post-attack 120 rounds, our attack success rate rose over 50% from baseline. The core code of our method is available at https://github.com/PhD-TaoLiu/FCBA.
Unlearnable example attacks are data poisoning attacks aiming to degrade the clean test accuracy of deep learning by adding imperceptible perturbations to the training samples, which can be formulated as a bi-level optimization problem. However, directly solving this optimization problem is intractable for deep neural networks. In this paper, we investigate unlearnable example attacks from a game-theoretic perspective, by formulating the attack as a nonzero sum Stackelberg game. First, the existence of game equilibria is proved under the normal setting and the adversarial training setting. It is shown that the game equilibrium gives the most powerful poison attack in that the victim has the lowest test accuracy among all networks within the same hypothesis space when certain loss functions are used. Second, we propose a novel attack method, called the Game Unlearnable Example (GUE), which has three main gradients. (1) The poisons are obtained by directly solving the equilibrium of the Stackelberg game with a first-order algorithm. (2) We employ an autoencoder-like generative network model as the poison attacker. (3) A novel payoff function is introduced to evaluate the performance of the poison. Comprehensive experiments demonstrate that GUE can effectively poison the model in various scenarios. Furthermore, the GUE still works by using a relatively small percentage of the training data to train the generator, and the poison generator can generalize to unseen data well. Our implementation code can be found at https://github.com/hong-xian/gue.
With the growth of 3D sensing technology, the deep learning system for 3D point clouds has become increasingly important, especially in applications such as autonomous vehicles where safety is a primary concern. However, there are growing concerns about the reliability of these systems when they encounter noisy point clouds, either occurring naturally or introduced with malicious intent. This paper highlights the challenges of point cloud classification posed by various forms of noise, from simple background noise to malicious adversarial/backdoor attacks that can intentionally skew model predictions. While there's an urgent need for optimized point cloud denoising, current point outlier removal approaches, an essential step for denoising, rely heavily on handcrafted strategies and are not adapted for higher-level tasks, such as classification. To address this issue, we introduce an innovative point outlier cleansing method that harnesses the power of downstream classification models. Using gradient-based attribution analysis, we define a novel concept: point risk. Drawing inspiration from tail risk minimization in finance, we recast the outlier removal process as an optimization problem, named PointCVaR. Extensive experiments show that our proposed technique not only robustly filters diverse point cloud outliers but also consistently and significantly enhances existing robust methods for point cloud classification. A notable feature of our approach is its effectiveness in defending against the latest threat of backdoor attacks in point clouds.
Revisiting the Information Capacity of Neural Network Watermarks: Upper Bound Estimation and Beyond
PDF ↗To trace the copyright of deep neural networks, an owner can embed its identity information into its model as a watermark. The capacity of the watermark quantify the maximal volume of information that can be verified from the watermarked model. Current studies on capacity focus on the ownership verification accuracy under ordinary removal attacks and fail to capture the relationship between robustness and fidelity. This paper studies the capacity of deep neural network watermarks from an information theoretical perspective. We propose a new definition of deep neural network watermark capacity analogous to channel capacity, analyze its properties, and design an algorithm that yields a tight estimation of its upper bound under adversarial overwriting. We also propose a universal non-invasive method to secure the transmission of the identity message beyond capacity by multiple rounds of ownership verification. Our observations provide evidence for neural network owners and defenders that are curious about the tradeoff between the integrity of their ownership and the performance degradation of their products.
Counterfactual explanations shed light on the decisions of black-box models by explaining how an input can be altered to obtain a favourable decision from the model (e.g., when a loan application has been rejected). However, as noted recently, counterfactual explainers may lack robustness in the sense that a minor change in the input can cause a major change in the explanation. This can cause confusion on the user side and open the door for adversarial attacks. In this paper, we study some sources of non-robustness. While there are fundamental reasons for why an explainer that returns a single counterfactual cannot be robust in all instances, we show that some interesting robustness guarantees can be given by reporting multiple rather than a single counterfactual. Unfortunately, the number of counterfactuals that need to be reported for the theoretical guarantees to hold can be prohibitively large. We therefore propose an approximation algorithm that uses a diversity criterion to select a feasible number of most relevant explanations and study its robustness empirically. Our experiments indicate that our method improves the state-of-the-art in generating robust explanations, while maintaining other desirable properties and providing competitive computational performance.
We examine machine learning models in a setup where individuals have the choice to share optional personal information with a decision-making system, as seen in modern insurance pricing models. Some users consent to their data being used whereas others object and keep their data undisclosed. In this work, we show that the decision not to share data can be considered as information in itself that should be protected to respect users' privacy. This observation raises the overlooked problem of how to ensure that users who protect their personal data do not suffer any disadvantages as a result. To address this problem, we formalize protection requirements for models which only use the information for which active user consent was obtained. This excludes implicit information contained in the decision to share data or not. We offer the first solution to this problem by proposing the notion of Protected User Consent (PUC), which we prove to be loss-optimal under our protection requirement. We observe that privacy and performance are not fundamentally at odds with each other and that it is possible for a decision maker to benefit from additional data while respecting users' consent. To learn PUC-compliant models, we devise a model-agnostic data augmentation strategy with finite sample convergence guarantees. Finally, we analyze the implications of PUC on challenging real datasets, tasks, and models.
Exponent Relaxation of Polynomial Zonotopes and Its Applications in Formal Neural Network Verification
PDF ↗Formal verification of neural networks is a challenging problem due to the complexity and nonlinearity of neural networks. It has been shown that polynomial zonotopes can tightly enclose the output set of a neural network. Unfortunately, the tight enclosure comes with additional complexity in the set representation, thus, rendering subsequent operations expensive to compute, such as computing interval bounds and intersection checking. To address this issue, we present a novel approach to restructure a polynomial zonotope to tightly enclose the original polynomial zonotope while drastically reducing its complexity. The restructuring is achieved by relaxing the exponents of the dependent factors of polynomial zonotopes and finding an appropriate approximation error. We demonstrate the applicability of our approach on output sets of neural networks, where we obtain tighter results in various subsequent operations, such as order reduction, zonotope enclosure, and range bounding.
Producing labels for unlabeled data is error-prone, making semi-supervised learning (SSL) troublesome. Often, little is known about when and why an algorithm fails to outperform a supervised baseline. Using benchmark datasets, we craft five common real-world SSL data scenarios: few-label, open-set, noisy-label, and class distribution imbalance/misalignment in the labeled and unlabeled sets. We propose a novel algorithm called Contrastive Credibility Propagation (CCP) for deep SSL via iterative transductive pseudo-label refinement. CCP unifies semi-supervised learning and noisy label learning for the goal of reliably outperforming a supervised baseline in any data scenario. Compared to prior methods which focus on a subset of scenarios, CCP uniquely outperforms the supervised baseline in all scenarios, supporting practitioners when the qualities of labeled or unlabeled data are unknown.
Randomized smoothing is currently the state-of-the-art method that provides certified robustness for deep neural networks. However, due to its excessively conservative nature, this method of incomplete verification often cannot achieve an adequate certified radius on real-world datasets. One way to obtain a larger certified radius is to use an input-specific algorithm instead of using a fixed Gaussian filter for all data points. Several methods based on this idea have been proposed, but they either suffer from high computational costs or gain marginal improvement in certified radius. In this work, we show that by exploiting the quasiconvex problem structure, we can find the optimal certified radii for most data points with slight computational overhead. This observation leads to an efficient and effective input-specific randomized smoothing algorithm. We conduct extensive experiments and empirical analysis on CIFAR-10 and ImageNet. The results show that the proposed method significantly enhances the certified radii with low computational overhead.
Partial observability and uncertainty are common problems in sequential decision-making that particularly impede the use of formal models such as Markov decision processes (MDPs). However, in practice, agents may be able to employ costly sensors to measure their environment and resolve partial observability by gathering information. Moreover, imprecise transition functions can capture model uncertainty. We combine these concepts and extend MDPs to robust active-measuring MDPs (RAM-MDPs). We present an active-measure heuristic to solve RAM-MDPs efficiently and show that model uncertainty can, counterintuitively, let agents take fewer measurements. We propose a method to counteract this behavior while only incurring a bounded additional cost. We empirically compare our methods to several baselines and show their superior scalability and performance.
Recent research has introduced several approaches to formally verify the robustness of neural network models against perturbations in their inputs, such as the ones that occur in adversarial attacks. At the same time, this particular verification task is known to be computationally challenging. More specifically, assessing the robustness of a neural network against input perturbations can easily take several hours of compute time per input vector, even when using state-of-the-art verification approaches. In light of this, it becomes challenging to select from a given set of neural network models the one that is best in terms of robust accuracy, i.e., the fraction of instances for which the model is known to be robust against adversarial perturbations, especially when given limited computing resources. To tackle this problem, we propose a racing method specifically adapted to the domain of robustness verification. This racing method utilises Delta-values, which can be seen as an efficiently computable proxy for the distance of a given input to a neural network model to the decision boundary. We present statistical evidence indicating significant differences in the empirical cumulative distribution between robust and non-robust inputs as a function of Delta-values. Using this information, we show that it is possible to reliably expose vulnerabilities in the model with relatively few input iterations. Overall, when applied to selecting the most robust network from sets of 31 MNIST and 27 CIFAR-10 networks, our proposed method achieves speedups of a factor of 108 and 42, respectively, in terms of cumulative running time compared to standard local robustness verification on the complete testing sets.