URL-based authentication provides a centralized and flexible way to safeguard sensitive resources in Java web applications by enforcing authentication checks based on URL paths. However, inconsistencies in handling flexible routing features (e.g., removing /./) between URL routing and authentication can be exploited to bypass authentication checks, resulting in URL-based Authentication Bypass Vulnerabilities (UABVulns). These vulnerabilities allow attackers to access sensitive resources without authentication, leading to serious security breaches. In this paper, we conduct the first in-depth study of 53 real-world UABVulns in Java web applications. Our study uncovers the root causes of UABVulns and identifies three key findings regarding URL routing, authentication, and sanitization. Guided by these findings, we design and implement UABScan, a static analysis tool that detects UABVulns by matching routing and authentication inconsistencies through pattern-based analysis. We evaluate UABScan on 529 popular Java web applications and successfully report 94 UABVulns across 72 applications, including 35 verified high-risk 0-days. Through manual investigation, UABScan achieves a recall of 87.50% and a precision of 80.00%, and significantly outperforms the state-of-the-art tool. To date, 31 CVE IDs have been assigned.

Buoyancy systems play a vital role in the efficient movement and control of deep-sea robots. Traditional buoyancy systems for these robots rely on high-pressure hydraulic pumps and heavy pressure-resistant shells, resulting in notable increases in size, mass, and cost. Consequently, there is an urgent need for a lightweight, high-pressure-resistant, self-sensing buoyancy adjustment device that can accommodate the multimodal movement requirements of miniaturized deep-sea robots. Drawing inspiration from the buoyancy regulation mechanism of sperm whales, we have developed a self-sensing phase-change buoyancy regulation system tailored for miniaturized robots, designed to operate in extreme deep-sea environments. Each module weighs only 35g while providing 4g of buoyancy adjustment. Experimental results demonstrate that the system maintains stable operation under hydrostatic pressures of 30MPa. When integrated into a miniaturized deep-sea robot, this module successfully enables controlled buoyancy for multimodal movement in aquatic environments.

Generating CT images from radiographs has immense clinical potential, offering a novel approach to low-cost and low-radiation medical imaging. This method can significantly ease the workload of radiologists. Current machine and deep learning approaches mainly utilize fully supervised learning methods for CT image generation tasks. However, the feature representations learned in fully supervised settings are typically task-specific, dependent on large quantities of labeled data, and have limited generalizability. Addressing these challenges, we introduce a contrastive learning-based method for CT image generation, which includes two phases: pre-training and fine-tuning. The pre-training phase aims to learn feature representations from unlabeled radiographs. Specifically, we devised two simple yet efficient radiograph data augmentation methods, converting the original data into two related but different views. These are then input into an encoder module to learn discriminative feature representations. Labeled data is used to learn the medical image generation task during the fine-tuning phase. In evaluations across the LIDC-IDRI lung CT and IXI brain MRI datasets, our CLCT-GAN model exhibits not only outstanding performance in lung CT reconstructions but also showcases remarkable adaptability to brain MRI data from IXI, surpassing previous state-of-the-art models in these diverse medical imaging benchmarks.