Mark Ryan

Endong Liu, Mark Ryan, Liyi Zhou et al.

Sanctioning blockchain addresses has become a common regulatory response to malicious activities. However, enforcement on permissionless blockchains remains challenging due to complex transaction flows and sophisticated fund-obfuscation techniques. Using cryptocurrency mixing tool Tornado Cash as a case study, we quantitatively assess the effectiveness of U.S. Office of Foreign Assets Control (OFAC) sanctions over a 957-day period, covering 6.79 million Ethereum blocks and 1.07 billion transactions. Our analysis reveals that while OFAC sanctions reduced overall Tornado Cash deposit volume by 71.03% to approximately 2 billion USD, attackers still relied on Tornado Cash in 78.33% of Ethereum-related security incidents, underscoring persistent evasion strategies. In this paper, we identify three significant, structural limitations in current sanction enforcement practices: (i) fragmented censorship in blockchain consensus and application layer; (ii) the complexity of obfuscation virtual asset services exploited by users; and (iii) the susceptibility of naive binary sanction classifications to dusting attacks. Our analysis and findings contribute to ongoing discussions around regulatory effectiveness in Decentralized Finance by providing empirical evidence, clarifying enforcement challenges, and informing future compliance strategies in response to sanctions and blockchain-based security risks.

Daniel Fentham, David Parker, Mark Ryan

When deploying classifiers in the real world, users expect them to respond to inputs appropriately. However, traditional classifiers are not equipped to handle inputs which lie far from the distribution they were trained on. Malicious actors can exploit this defect by making adversarial perturbations designed to cause the classifier to give an incorrect output. Classification-with-rejection methods attempt to solve this problem by allowing networks to refuse to classify an input in which they have low confidence. This works well for strongly adversarial examples, but also leads to the rejection of weakly perturbed images, which intuitively could be correctly classified. To address these issues, we propose Reed-Muller Aggregation Networks (RMAggNet), a classifier inspired by Reed-Muller error-correction codes which can correct and reject inputs. This paper shows that RMAggNet can minimise incorrectness while maintaining good correctness over multiple adversarial attacks at different perturbation budgets by leveraging the ability to correct errors in the classification process. This provides an alternative classification-with-rejection method which can reduce the amount of additional processing in situations where a small number of incorrect classifications are permissible.

Jesse Spielman, David Oswald, Mark Ryan et al.

With high-stakes machine learning applications increasingly moving to untrusted end-user or cloud environments, safeguarding pre-trained model parameters becomes essential for protecting intellectual property and user privacy. Recent advancements in hardware-isolated enclaves, notably Intel SGX, hold the promise to secure the internal state of machine learning applications even against compromised operating systems. However, we show that privileged software adversaries can exploit input-dependent memory access patterns in common neural network activation functions to extract secret weights and biases from an SGX enclave. Our attack leverages the SGX-Step framework to obtain a noise-free, instruction-granular page-access trace. In a case study of an 11-input regression network using the Tensorflow Microlite library, we demonstrate complete recovery of all first-layer weights and biases, as well as partial recovery of parameters from deeper layers under specific conditions. Our novel attack technique requires only 20 queries per input per weight to obtain all first-layer weights and biases with an average absolute error of less than 1%, improving over prior model stealing attacks. Additionally, a broader ecosystem analysis reveals the widespread use of activation functions with input-dependent memory access patterns in popular machine learning frameworks (either directly or via underlying math libraries). Our findings highlight the limitations of deploying confidential models in SGX enclaves and emphasise the need for stricter side-channel validation of machine learning implementations, akin to the vetting efforts applied to secure cryptographic libraries.

Mihai Ordean, Mark Ryan, David Galindo

This paper proposes Concurrent-Access Obfuscated Store (CAOS), a construction for remote data storage that provides access-pattern obfuscation in a honest-but-curious adversarial model, while allowing for low bandwidth overhead and client storage. Compared to the state of the art, the main advantage of CAOS is that it supports concurrent access without a proxy, for multiple read-only clients and a single read-write client. Concurrent access is achieved by letting clients maintain independent maps that describe how the data is stored. These maps might diverge from client to client, but it is guaranteed that no client will ever lose track of current data. We achieve efficiency and concurrency at the expense of perfect obfuscation: in CAOS the extent to which access patterns are hidden is determined by the resources allocated to its built-in obfuscation mechanism. To assess this trade-off we provide both a security and a performance analysis of our protocol instance. We additionally provide a proof-of-concept implementation.

Jiangshan Yu, Vincent Cheval, Mark Ryan

The security of public key validation protocols for web-based applications has recently attracted attention because of weaknesses in the certificate authority model, and consequent attacks. Recent proposals using public logs have succeeded in making certificate management more transparent and verifiable. However, those proposals involve a fixed set of authorities. This means an oligopoly is created. Another problem with current log-based system is their heavy reliance on trusted parties that monitor the logs. We propose a distributed transparent key infrastructure (DTKI), which greatly reduces the oligopoly of service providers and allows verification of the behaviour of trusted parties. In addition, this paper formalises the public log data structure and provides a formal analysis of the security that DTKI guarantees.

Masoud Koleini, Eike Ritter, Mark Ryan

We develop a modeling technique based on interpreted systems in order to verify temporal-epistemic properties over access control policies. This approach enables us to detect information flow vulnerabilities in dynamic policies by verifying the knowledge of the agents gained by both reading and reasoning about system information. To overcome the practical limitations of state explosion in model-checking temporal-epistemic properties, we introduce a novel abstraction and refinement technique for temporal-epistemic safety properties in ACTLK (ACTL with knowledge modality K) and a class of interesting properties that does fall in this category.