Towards Provable (In)Secure Model Weight Release Schemes

Recent secure weight release schemes claim to enable open-source model distribution while protecting model ownership and preventing misuse. However, these approaches lack rigorous security foundations and provide only informal security guarantees. Inspired by established works in cryptography, we...

Bridging Unsupervised and Semi-Supervised Anomaly Detection: a Theoretically-Grounded and Practical Framework with Synthetic Anomalies

Anomaly detection AD is a critical task across domains such as cybersecurity and healthcare. In the unsupervised setting, an effective and theoretically-grounded principle is to train classifiers to distinguish normal data from synthetic anomalies. We extend this principle to semi-supervised AD,...

Certified Unlearning for Neural Networks

We address the problem of machine unlearning, where the goal is to remove the influence of specific training data from a model upon request, motivated by privacy concerns and regulatory requirements such as the "right to be forgotten." Unfortunately, existing methods rely on restrictive assumptio...

A Certified Unlearning Approach without Access to Source Data

With the growing adoption of data privacy regulations, the ability to erase private or copyrighted information from trained models has become a crucial requirement. Traditional unlearning methods often assume access to the complete training dataset, which is unrealistic in scenarios where the...

GeoClip: Geometry-Aware Clipping for Differentially Private SGD

Differentially private stochastic gradient descent DP-SGD is the most widely used method for training machine learning models with provable privacy guarantees. A key challenge in DP-SGD is setting the per-sample gradient clipping threshold, which significantly affects the trade-off between privac...

Breaking the Gaussian Barrier: Residual-PAC Privacy for Automatic Privatization

The Probably Approximately Correct PAC Privacy framework 1 provides a powerful instance-based methodology for certifying privacy in complex data-driven systems. However, existing PAC Privacy algorithms rely on a Gaussian mutual information upper bound. We show that this is in general too...

QA-HFL: Quality-Aware Hierarchical Federated Learning for Resource-Constrained Mobile Devices with Heterogeneous Image Quality

This paper introduces QA-HFL, a quality-aware hierarchical federated learning framework that efficiently handles heterogeneous image quality across resource-constrained mobile devices. Our approach trains specialized local models for different image quality levels and aggregates their features...

CVE-2025-48946

CVE-2025-48946 concerns the liboqs library (C), specifically the HQC algorithm implemented in versions prior to 0.13.0. The root cause is a theoretical design flaw in HQC that can lead to large numbers of malformed ciphertexts sharing the same implicit rejection value. The public descriptions sta...

CVE-2025-48946 liboqs affected by theoretical design flaw in HQC

liboqs is a C-language cryptographic library that provides implementations of post-quantum cryptography algorithms. liboqs prior to version 0.13.0 supports the HQC algorithm, an algorithm with a theoretical design flaw which leads to large numbers of malformed ciphertexts sharing the same implici...

CVE-2025-48946 liboqs affected by theoretical design flaw in HQC

liboqs is a C-language cryptographic library that provides implementations of post-quantum cryptography algorithms. liboqs prior to version 0.13.0 supports the HQC algorithm, an algorithm with a theoretical design flaw which leads to large numbers of malformed ciphertexts sharing the same implici...

Engineering Trustworthy Machine-Learning Operations with Zero-Knowledge Proofs

As Artificial Intelligence AI systems, particularly those based on machine learning ML, become integral to high-stakes applications, their probabilistic and opaque nature poses significant challenges to traditional verification and validation methods. These challenges are exacerbated in regulated...

Transaction Fee Mechanism Design for Leaderless Blockchain Protocols

We initiate the study of transaction fee mechanism design for blockchain protocols in which multiple block producers contribute to the production of each block. Our contributions include: - We propose an extensive-form multi-stage game model to reason about the game theory of multi-proposer...

CVE-2019-19391

In LuaJIT through 2.0.5, as used in Moonjit before 2.1.2 and other products, debug.getinfo has a type confusion issue that leads to arbitrary memory write or read operations, because certain cases involving valid stack levels and options are mishandled. NOTE: The LuaJIT project owner states that...

PRUNE: a Patching Based Repair Framework for Certifiable Unlearning of Neural Networks

It is often desirable to remove a.k.a. unlearn a specific part of the training data from a trained neural network model. A typical application scenario is to protect the data holder's right to be forgotten, which has been promoted by many recent regulation rules. Existing unlearning methods invol...

Engineering Risk-Aware, Security-By-Design Frameworks for Assurance of Large-Scale Autonomous AI Models

As AI models scale to billions of parameters and operate with increasing autonomy, ensuring their safe, reliable operation demands engineering-grade security and assurance frameworks. This paper presents an enterprise-level, risk-aware, security-by-design approach for large-scale autonomous AI...

Bilateral Differentially Private Vertical Federated Boosted Decision Trees

Federated learning is a distributed machine learning paradigm that enables collaborative training across multiple parties while ensuring data privacy. Gradient Boosting Decision Trees GBDT, such as XGBoost, have gained popularity due to their high performance and strong interpretability. Therefor...

WhatsApp Launches Private Processing to Enable AI Features While Protecting Message Privacy

Popular messaging app WhatsApp on Tuesday unveiled a new technology called Private Processing to enable artificial intelligence AI capabilities in a privacy-preserving manner. "Private Processing will allow users to leverage powerful optional AI features – like summarizing unread messages or...

SILENT: a New Lens on Statistics in Software Timing Side Channels

Cryptographic research takes software timing side channels seriously. Approaches to mitigate them include constant-time coding and techniques to enforce such practices. However, recent attacks like Meltdown 42, Spectre 37, and Hertzbleed 70 have challenged our understanding of what it means for...

CVE-2025-27508 Emissary Use of a Broken or Risky Cryptographic Algorithm

Emissary is a P2P based data-driven workflow engine. The ChecksumCalculator class within allows for hashing and checksum generation, but it includes or defaults to algorithms that are no longer recommended for secure cryptographic use cases e.g., SHA-1, CRC32, and SSDEEP. These algorithms, while...

CVE-2025-27508

Emissary (CVE-2025-27508) uses a ChecksumCalculator that defaults to weak cryptographic algorithms (SHA-1, CRC32, SSDEEP). The issue may undermine cryptographic guarantees in contexts requiring strong integrity, with potential collision risks or tampering awareness. The problem is documented acro...