Algorithm ensures privacy while strictly adhering to constraints.
problem Differential privacy with linear constraints that must be strictly followed.
method Developed an algorithm that releases a nearly-optimal solution satisfying constraints with probability 1.
result Achieved nearly optimal performance while preserving privacy and strictly adhering to constraints.
Privacy concerns have led to the development of privacy-preserving approaches for learning models from sensitive data. Yet, in practice, even models learned with privacy guarantees can inadvertently memorize unique training examples or leak sensitive features. To identify such privacy violations, existing model auditin…
Our everyday interactions with pervasive systems generate traces that capture various aspects of human behavior and enable machine learning algorithms to extract latent information about users. In this paper, we propose a machine learning interpretability framework that enables users to understand how these generated t…
Differential privacy improves AI security, fairness, and learning.
problem Privacy violations, security issues, and model fairness in AI.
method Application of differential privacy in various AI areas.
result Differential privacy enhances AI performance in multiple areas.
Language models learn from training data and can leak private information.
problem Language models lack context understanding and can expose private data.
method Discussing the limitations of current privacy protection methods for language models.
result Existing privacy protection methods are insufficient for language models.
New approach protects privacy of deleted records in machine learning.
problem Privacy of deleted records in machine learning models.
method Sound deletion guarantee and noisy gradient descent algorithm.
result Privacy of existing records is necessary for deleted records' privacy.
Deep learning models are often trained on datasets that contain sensitive information such as individuals' shopping transactions, personal contacts, and medical records. An increasingly important line of work therefore has sought to train neural networks subject to privacy constraints that are specified by differential…
New auditors assess f-DP privacy with adaptive sampling, avoiding large sample sizes.
problem Empirical auditing of f-DP privacy with adaptive sampling. method Shift focus to f-DP, develop adaptive auditors for whitebox and blackbox settings. result Adaptive auditors detect f-DP violations across the privacy spectrum with statistical guarantees. Decor protects decentralized learning models from curious users.
problem Privacy violation in decentralized learning.
method Decor uses correlated Gaussian noises to protect local models in decentralized SGD with differential privacy guarantees.
result Decor matches central DP optimal privacy-utility trade-off for arbitrary connected graphs.
In distributed reinforcement learning, it is common to exchange the experience memory of each agent and thereby collectively train their local models. The experience memory, however, contains all the preceding state observations and their corresponding policies of the host agent, which may violate the privacy of the ag…
This paper addresses privacy issues in personalized pricing using nonparametric demand models.
problem Privacy violation in personalized pricing algorithms with unknown nonparametric demand models.
method Develops algorithms to make pricing decisions and learn demand while ensuring central and local differential privacy.
result Proves near-optimal regret bounds for algorithms with CDP and LDP guarantees.
FSL-BDP models time-to-default without centralizing data, improving privacy mechanisms in federated settings.
problem Traditional credit risk models ignore default timing and violate data-protection rules.
method Federated Survival Learning with Bayesian Differential Privacy (FSL-BDP).
result FSL-BDP improves privacy mechanisms in federated settings, outperforming classical DP in most clients.
Deep Neural Network (DNN) workloads are quickly moving from datacenters onto edge devices, for latency, privacy, or energy reasons. While datacenter networks can be protected using conventional cybersecurity measures, edge neural networks bring a host of new security challenges. Unlike classic IoT applications, edge ne…
"Overlearning" means that a model trained for a seemingly simple objective implicitly learns to recognize attributes and concepts that are (1) not part of the learning objective, and (2) sensitive from a privacy or bias perspective. For example, a binary gender classifier of facial images also learns to recognize races…
This paper enhances privacy in statistical model checking of cyber-physical systems.
problem Privacy concerns in consumer-level applications due to statistical model checking.
method Proposes expected differential privacy and a new exponential mechanism for sequential algorithms.
result Demonstrates a novel mechanism to preserve privacy in statistical model checking.
FedOS tackles challenges in federated learning by using open-set learning.
problem Challenges in federated learning due to data locality and privacy constraints.
method Introduces open-set learning to stabilize training in federated learning.
result Demonstrates improved model performance through open-set learning.
New method reconstructs data subsets from limited published statistics.
problem Reconstructing tabular data from aggregate statistics when full datasets are not possible.
method Generates and verifies subsets of rows and columns that are guaranteed to be correct.
result Privacy violations can persist even with sparse published statistics.
New method for private learning with fairness constraints.
problem Rate-constrained optimization under differential privacy.
method RaCO-DP, a DP variant of SGDA solving Lagrangian formulation.
result Empirical results show RaCO-DP outperforms existing methods.
The protection of user privacy is an important concern in machine learning, as evidenced by the rolling out of the General Data Protection Regulation (GDPR) in the European Union (EU) in May 2018. The GDPR is designed to give users more control over their personal data, which motivates us to explore machine learning fr…
Efficiently estimates prediction error in regression with Gaussian covariates under privacy constraints.
problem Private regression with Gaussian covariates under differential privacy constraints.
method Sum-of-Squares framework combined with robust estimators.
result Sample-optimal private regression algorithm with optimal error rates.
The scarcity of data and isolated data islands encourage different organizations to share data with each other to train machine learning models. However, there are increasing concerns on the problems of data privacy and security, which urges people to seek a solution like Federated Transfer Learning (FTL) to share trai…
Federated CycleGAN enables privacy-preserving image translation without central data.
problem Privacy and security issues in unsupervised image-to-image translation.
method Novel federated CycleGAN architecture that decomposes CycleGAN loss into client-specific local objectives.
result Federated CycleGAN achieves comparable performance to non-federated CycleGAN without central data exchange.
Proactive DP optimizes privacy and utility in DP-SGD with a fixed privacy budget.
problem Balancing privacy and utility in differential privacy for machine learning.
method Proposes a pro-active DP framework that allows a-priori selection of DP-SGD parameters to maximize test accuracy.
result Proactive DP can optimize utility of DP-SGD with a fixed privacy budget (ε, δ).
Framework audits synthetic datasets for trustworthiness across various use cases.
problem Assessing the trustworthiness of synthetic datasets and models.
method Holistic auditing framework focusing on bias, fidelity, utility, robustness, and privacy.
result Introduces a trustworthiness index and model selection process for controllable trade-offs.
Paper relaxes SGD privacy and generalization guarantees for non-smooth convex losses.
problem Privacy and generalization in SGD for non-smooth convex losses.
method Relaxes Lipschitz and strong smoothness assumptions to Hölder smoothness, proving (ε,δ)-DP and optimal excess risk. result Noisy SGD with α-Hölder smooth losses achieves optimal excess risk with linear gradient complexity for α≥1/2. Face recognition systems are vulnerable to composite face reconstruction attacks.
problem Vulnerability of face recognition systems to composite face reconstruction attacks.
method Assumed attacker uses composite face parts to reconstruct faces faster and more efficiently.
result Current face recognition systems are extremely vulnerable to random search attacks.
Federated Learning prioritizes client data contributions for better model quality.
problem Privacy concerns and reluctance to share private data in machine learning.
method Prioritizes client data contributions in Federated Learning by assigning scores based on defined criteria.
result The proposed approach yields a higher quality global model compared to standard Federated Learning.
WAFFLe anonymizes federated learning weights to protect data privacy and fairness.
problem Federated learning exposes local models to attacks and underfits heterogeneous clients.
method Combines Indian Buffet Process with shared weight factors.
result Significant improvement in local test performance and fairness.
New approach shows data memorization trade-offs in large models.
problem Data memorization in large language models and its privacy implications.
method Developed a new approach using strong data processing inequalities to prove lower bounds on memorization.
result Proved that Ω(d) bits of training data information must be memorized for O(1) examples, decaying with example growth. Bayesian methods detect significant IIA violations in similarity choice data.
problem Detecting IIA violations in similarity choice data complicates classical models.
method Proposed two statistical methods: classical goodness-of-fit test and Bayesian PPC.
result Significant IIA violations confirmed in both datasets, driven by context effects.
This paper examines how Higher-Order Langevin Dynamics reduces memorization in diffusion models.
problem Memorization of training samples in diffusion models, violating copyright and privacy.
method Introduces Higher-Order Langevin Dynamics (HOLD) to regularize diffusion model trajectories.
result The dynamics of the data variable in HOLD are governed by a low-pass-filtered version of the learned score function, with smoothness increasing with model order.
The assumption of positivity in causal inference (also known as common support and co-variate overlap) is necessary to obtain valid causal estimates. Therefore, confirming it holds in a given dataset is an important first step of any causal analysis. Most common methods to date are insufficient for discovering non-posi…
The correspondence between Riemann-Finsler geometries and effective field theories with spin-independent Lorentz violation is explored. We obtain the general quadratic action for effective scalar field theories in any spacetime dimension with Lorentz-violating operators of arbitrary mass dimension. Classical relativist…
Non-affine aggregation rules cannot preserve monotonicity in convex learning.
problem Designing non-affine aggregation rules that maintain monotonicity in convex learning.
method Proving that monotonicity of aggregated gradients is preserved only if the aggregation rule is positively affine.
result Non-affine aggregation prevents steady convergence and substantially degrades algorithmic stability.
This contribution to the CPT'13 meeting briefly introduces Lorentz and CPT violation and outlines two recent developments in the field.
Bipartite Riemann-Finsler geometries with complementary Finsler structures are constructed. Calculable examples are presented based on a bilinear-form coefficient for explicit Lorentz violation.
Detecting faults and SLA violations in a timely manner is critical for telecom providers, in order to avoid loss in business, revenue and reputation. At the same time predicting SLA violations for user services in telecom environments is difficult, due to time-varying user demands and infrastructure load conditions. In…
We present a new approach to assessing the robustness of neural networks based on estimating the proportion of inputs for which a property is violated. Specifically, we estimate the probability of the event that the property is violated under an input model. Our approach critically varies from the formal verification f…
CVTMLE improves statistical inference in settings of positivity or Donsker class violations.
problem Inference issues in causal inference due to data sparsity or near-positivity violations.
method Cross-validation of TMLE (CVTMLE) to improve performance in settings of positivity or Donsker class violations.
result CVTMLE vastly improves confidence interval coverage without affecting bias, especially in small sample sizes and near-positivity violations.
The study proves strong cosmic censorship violation for spherically symmetric dust clouds.
problem Violation of strong cosmic censorship for spherically symmetric dust clouds.
method Derived an ordinary differential equation for light rays and used it to prove strong cosmic censorship violation.
result Generic violation of strong cosmic censorship for spherically symmetric dust clouds.
Improves privacy amplification by shuffling for differential privacy.
problem Enhancing privacy guarantees in systems with anonymous data contributions.
method Theoretical and numerical analysis of Rényi differential privacy parameters and privacy amplification by shuffling.
result First asymptotically optimal analysis of Rényi differential privacy parameters for shuffled outputs.
New algorithm reduces sample complexity for safe reinforcement learning.
problem Safe reinforcement learning in constrained MDPs with performance and safety constraints.
method Model-based primal-dual algorithm balancing regret and bounded constraint violations.
result Proves near-optimal policies with bounded violations or zero violations in CMDPs.
Proposes a privacy-preserving recommendation system using matrix factorization and differential privacy.
problem Privacy leakage in recommendation systems when anonymizing user data is not sufficient.
method Uses matrix factorization and differential privacy via the Gaussian mechanism.
result Demonstrates excellent utility for privacy-preserving recommendation systems.
We show that any open subset of a contact manifold of dimension greater than three contains a certain non-convex hypersurface violating the Thurston-Bennequin inequality.
Recommender systems are widely used to predict personalized preferences of goods or services using users' past activities, such as item ratings or purchase histories. If collections of such personal activities were made publicly available, they could be used to personalize a diverse range of services, including targete…
This work optimizes mean estimation under varying privacy constraints.
problem Mean estimation with heterogeneous privacy constraints.
method Proposes an algorithm for mean estimation under different privacy levels for users.
result Shows a saturation phenomenon in performance as privacy levels are relaxed.
Improves privacy guarantees by analyzing randomness in privacy-preserving mechanisms.
problem Balancing user privacy and business constraints in privacy-preserving mechanisms.
method Analyzes explicit and implicit randomness in privacy mechanisms and proposes a probabilistic calibration method.
result Proposes privacy at risk, providing stronger privacy guarantees with quantifiable risks.
This paper considers online convex optimization over a complicated constraint set, which typically consists of multiple functional constraints and a set constraint. The conventional online projection algorithm (Zinkevich, 2003) can be difficult to implement due to the potentially high computation complexity of the proj…