Related papers: How Query Distribution Knowledge Breaks Multidimen…
Standard quantum key distribution protocols are provably secure against eavesdropping attacks, if quantum theory is correct. It is theoretically interesting to know if we need to assume the validity of quantum theory to prove the security…
Side-channel attacks currently constitute the main challenge for quantum key distribution (QKD) to bridge theory with practice. So far two main approaches have been introduced to address this problem, (full) device-independent QKD and…
Federated Learning (FL) trains a black-box and high-dimensional model among different clients by exchanging parameters instead of direct data sharing, which mitigates the privacy leak incurred by machine learning. However, FL still suffers…
Large Language Models (LLMs) are increasingly served on shared accelerators where an adversary with read access to device memory can observe KV caches and hidden states, threatening prompt privacy for open-source models. Cryptographic…
Sequence models, such as Large Language Models (LLMs) and autoregressive image generators, have a tendency to memorize and inadvertently leak sensitive information. While this tendency has critical legal implications, existing tools are…
Large Language Model (LLM) agents have achieved rapid adoption and demonstrated remarkable capabilities across a wide range of applications. To improve reasoning and task execution, modern LLM agents would incorporate memory modules or…
Randomized response is a popular local anonymization approach that can deliver anonymized multi-dimensional data sets with rigorous privacy guarantees. At the same time, it can ensure validity for exploratory analysis and machine learning…
Quantum key distribution (QKD) is a concept of secret key exchange supported by fundamentals of quantum physics. Its perfect realization offers unconditional key security, however, known practical schemes are potentially vulnerable if the…
Key distribution plays a fundamental role in cryptography. Currently, the quantum scheme stands as the only known method for achieving unconditionally secure key distribution. This method has been demonstrated over distances of 508 and 1002…
Most security proofs of quantum key distribution (QKD) assume that there is no unwanted information leakage about the state preparation process. However, this assumption is impossible to guarantee in practice, as QKD systems can leak…
A distribution inference attack aims to infer statistical properties of data used to train machine learning models. These attacks are sometimes surprisingly potent, but the factors that impact distribution inference risk are not well…
Plaintext non-delayed chaotic cipher (PNDCC) means that in the diffusion equation, plaintext has no delay terms while ciphertext has a feedback term. In existing literature, chaotic cipher diffusions invariably take this form. Since its…
In the digital age, image encryption technology acts as a safeguard, preventing unauthorized access to images. This paper proposes an innovative image encryption scheme that integrates a novel 2D hyper-chaotic map with a newly developed…
The distributed (federated) LLM is an important method for co-training the domain-specific LLM using siloed data. However, maliciously stealing model parameters and data from the server or client side has become an urgent problem to be…
With the rise of quantum technologies, data security increasingly relies on quantum cryptography and its most notable application, quantum key distribution (QKD). Yet, current technological limitations, in particular, the unavailability of…
While large language models (LLMs) perform strongly on diverse tasks, their trustworthiness is limited by erratic behavior that is unfaithful to their internal knowledge. In particular, LLMs often fail on multiple-choice questions (MCQs)…
Diffusion language models (DLMs) have recently emerged as a promising alternative to autoregressive (AR) approaches, enabling parallel token generation beyond a rigid left-to-right order. Despite growing empirical success, the theoretical…
In this work we present a security analysis for quantum key distribution, establishing a rigorous tradeoff between various protocol and security parameters for a class of entanglement-based and prepare-and-measure protocols. The goal of…
Mistrustful quantum cryptographic protocols encode information in incompatible observables, so that any attempt by a dishonest party to access multiple pieces of information necessarily involves a tradeoff. A natural class of such…
Counting the fraction of a population having an input within a specified interval i.e. a \emph{range query}, is a fundamental data analysis primitive. Range queries can also be used to compute other interesting statistics such as…