Synthetic aperture radar (SAR) images are often contaminated by speckle noise, a type of multiplicative noise resulting from the imaging process. SAR image despeckling is a crucial preprocessing step for satellite imaging, enhancing image visualization and facilitating downstream analysis. In this study, we propose a linear–angular attention transformer network for SAR despeckling. The approach efficiently captures both local and global context in linear time within a multiscale transformer-convolutional neural network architecture. Our transformer integrates nonlocal denoising and multiscale feature extraction in a single model. Using a smoothing loss and fast nonlocal post-processing, the model achieved a 17% improvement in structural similarity index, a 61% enhancement in contrast-to-noise ratio, and an increase above 100% in the equivalent number of looks metric across three datasets compared to multiple state-of-the-art baselines, even when trained on a small dataset and for only 13 epochs. Comparison with different multi-head self-attention mechanisms revealed the effectiveness of linear–angular attention as a step towards green AI, showing both quantitative and qualitative performance improvements. Unlike models that rely on optical images for training and lack domain-specific features for real SAR despeckling, the proposed network is trained directly on SAR images in a self-supervised manner.

As large language models (LLMs) become increasingly prevalent across many real-world applications, understanding and enhancing their robustness to adversarial attacks is of paramount importance. Existing methods for identifying adversarial prompts tend to focus on specific domains, lack diversity, or require extensive human annotations. To address these limitations, we present Rainbow Teaming, a novel black-box approach for producing a diverse collection of adversarial prompts. Rainbow Teaming casts adversarial prompt generation as a quality-diversity problem and uses open-ended search to generate prompts that are both effective and diverse. Focusing on the safety domain, we use Rainbow Teaming to target various state-of-the-art LLMs, including the Llama 2 and Llama 3 models. Our approach reveals hundreds of effective adversarial prompts, with an attack success rate exceeding 90% across all tested models. Furthermore, we demonstrate that prompts generated by Rainbow Teaming are highly transferable and that fine-tuning models with synthetic data generated by our method significantly enhances their safety without sacrificing general performance or helpfulness. We additionally explore the versatility of Rainbow Teaming by applying it to question answering and cybersecurity, showcasing its potential to drive robust open-ended self-improvement in a wide range of applications.

Historically, hands-on cybersecurity exercises helped reinforce the basic cybersecurity concepts. However, most of them focused on the user level attacks and defenses and did not provide a convenient way of studying the kernel level security. Since OS kernels provide foundations for applications, any compromise to OS kernels will lead to a computer that cannot be trusted. Moreover, there has been a great interest in using virtualization to profile, characterize, and observe kernel events including security incidents. Virtual Machine Introspection (VMI) is a technique that has been deeply investigated in intrusion detection, malware analysis, and memory forensics. Inspired by the great success of VMI, we used it to develop hands-on labs for teaching kernel level security. In this work, we present three VMI-based labs on (1) stack-based buffer over-flow, (2) direct kernel object manipulation (DKOM), and (3) kernel integrity checker which have been made available online. Then, we analyze the differences in approaches taken by VMI-based labs and traditional labs and conclude that VMI-based labs are better as opposed to traditional labs from a teaching standpoint because they provide more visibility than the traditional labs and superior ability to manipulate kernel memory which provides more insight into kernel security concepts.