arXiv subjects

Zhaoqian Liu

Publications and source records attributed to Zhaoqian Liu.

3 recordsLinked to original sources

arXiv2026

Diffusion-Aided Bandwidth-Efficient Semantic Communication with Adaptive Requests

Semantic communication focuses on conveying the task-relevant meaning rather than exact bitwise recovery. For image transmission with a generative receiver, relying only on text descriptions can be insufficient to preserve instance-specific visual evidence, whereas sending dense latent representations can incur substantial overhead. This paper presents a receiver-driven closed-loop scheme that transmits a short caption together with an initial sparse subset of latent blocks, and then uses feedback to request additional blocks only when needed. At each round, the receiver reconstructs the image via latent diffusion inpainting and applies a semantic consistency check between a caption generated from the reconstruction and the received caption, using a lightweight language similarity score such as ROUGE-L. The receiver stops early once a target consistency level is met, and otherwise requests a small number of additional latent blocks to refine the reconstruction. Experiments on Flickr30k over AWGN channels demonstrate a controllable rate-quality tradeoff. Adaptive feedback achieves the strongest semantic alignment and the lowest failure rate, outperforming budget-matched one-shot transmission while typically using fewer latent blocks than always-on retransmission.

cs.IT↗

arXiv2025

Diffusion-aided Task-oriented Semantic Communications with Model Inversion Attack

Semantic communication enhances transmission efficiency by conveying semantic information rather than raw input symbol sequences. Task-oriented semantic communication is a variant that tries to retains only task-specific information, thus achieving greater bandwidth savings. However, these neural-based communication systems are vulnerable to model inversion attacks, where adversaries try to infer sensitive input information from eavesdropped transmitted data. The key challenge, therefore, lies in preserving privacy while ensuring transmission correctness and robustness. While prior studies typically assume that adversaries aim to fully reconstruct the raw input in task-oriented settings, there exist scenarios where pixel-level metrics such as PSNR or SSIM are low, yet the adversary's outputs still suffice to accomplish the downstream task, indicating leakage of sensitive information. We therefore adopt the attacker's task accuracy as a more appropriate metric for evaluating attack effectiveness. To optimize the gap between the legitimate receiver's accuracy and the adversary's accuracy, we propose DiffSem, a diffusion-aided framework for task-oriented semantic communication. DiffSem integrates a transmitter-side self-noising mechanism that adaptively regulates semantic content while compensating for channel noise, and a receiver-side diffusion U-Net that enhances task performance and can be optionally strengthened by self-referential label embeddings. Our experiments demonstrate that DiffSem enables the legitimate receiver to achieve higher accuracy, thereby validating the superior performance of the proposed framework.

cs.CR↗

arXiv2025

Efficient and High-Accuracy Secure Two-Party Protocols for a Class of Functions with Real-number Inputs

In two-party secret sharing scheme, values are typically encoded as unsigned integers $\mathsf{uint}(x)$, whereas real-world applications often require computations on signed real numbers $\mathsf{Real}(x)$. To enable secure evaluation of practical functions, it is essential to computing $\mathsf{Real}(x)$ from shared inputs, as protocols take shares as input. At USENIX'25, Guo et al. proposed an efficient method for computing signed integer values $\mathsf{int}(x)$ from shares, which can be extended to compute $\mathsf{Real}(x)$. However, their approach imposes a restrictive input constraint $|x| < \frac{L}{3}$ for $x \in \mathbb{Z}_L$, limiting its applicability in real-world scenarios. In this work, we significantly relax this constraint to $|x| < B$ for any $B \leq \frac{L}{2}$, where $B = \frac{L}{2}$ corresponding to the natural representable range in $x \in \mathbb{Z}_L$. This relaxes the restrictions and enables the computation of $\mathsf{Real}(x)$ with loose or no input constraints. Building upon this foundation, we present a generalized framework for designing secure protocols for a broad class of functions, including integer division ($\lfloor \frac{x}{d} \rfloor$), trigonometric ($\sin(x)$) and exponential ($e^{-x}$) functions. Our experimental evaluation demonstrates that the proposed protocols achieve both high efficiency and high accuracy. Notably, our protocol for evaluating $e^{-x}$ reduces communication costs to approximately 31% of those in SirNN (S&P 21) and Bolt (S&P 24), with runtime speedups of up to $5.53 \times$ and $3.09 \times$, respectively. In terms of accuracy, our protocol achieves a maximum ULP error of $1.435$, compared to $2.64$ for SirNN and $8.681$ for Bolt.

cs.CR↗