Giandomenico Solimando

Text-to-image models (T2I) have demonstrated remarkable capabilities in generating realistic and high-quality images, finding applications in fields ranging from entertainment to social networks. However, the ease of access to these tools has raised concerns regarding their potential misuse from ethical and legal perspectives due to the possibility of creating disinformative, manipulative, or offensive content. In this paper, we investigate the effectiveness of the security systems of emerging T2I models to avoid generating harmful media using textual prompts. To this end, we introduce a new jailbreaking technique combining Projected Gradient Descent (PGD) attack with a safety-aware prompt optimization specifically tailored to bypass safety filters in advanced T2I models. Moreover, we propose a new defense mechanism inspired by the splitting of transformers’ attention layers of CLIP, leading to the definition of two classifiers for detecting unsafe and adversarial prompts. The evaluation of the proposed jailbreak approach highlighted the vulnerabilities of the most recent proprietary models, such as DALL·E, IMAGEN, and Nano Banana 2, allowing us to generate thousands of unsafe and unethical images related to seven harmful categories, i.e., harassment, hate, sexual, shocking, violence, self-harm, and illegal activities. Moreover, we have demonstrated that our approach outperforms most popular jailbreaking approaches in bypassing security filters and generating harmful images with an overall performance increase ranging from 2 to 8 times. The results of the proposed defense mechanism have shown that our mechanism is very effective in discriminating unsafe and adversarial prompts, outperforming the detection performance of the most popular commercial safety filters by up to 4 times, even against other types of adversarial prompts. The images and prompts were collected into three new datasets made available for benchmarking safety filters and analyzing the risks associated with T2I generation. Finally, we evaluate the performance of an ad-hoc classifier in detecting and identifying harmful content, achieving similar results to GPT-4o and Gemini 1.5 Pro in detection and enhanced performance in identification.

Although passwords remain the primary defense against unauthorized access, users often tend to use passwords that are easy to remember. This behavior significantly increases security risks, also due to the fact that traditional password strength evaluation methods are often inadequate. In this discussion paper, we present SODA ADVANCE, a data reconstruction tool also designed to enhance evaluation processes related to the password strength. In particular, SODA ADVANCE integrates a specialized module aimed at evaluating password strength by leveraging publicly available data from multiple sources, including social media platforms. Moreover, we investigate the capabilities and risks associated with emerging Large Language Models (LLMs) in evaluating and generating passwords, respectively. Experimental assessments conducted with 100 real users demonstrate that LLMs can generate strong and personalized passwords possibly defined according to user profiles. Additionally, LLMs were shown to be effective in evaluating passwords, especially when they can take into account user profile data.

The spatial-temporal data of future freight traffic speed in the metropolitan region must be properly understood to develop freight-related traffic management strategies. This work introduces a new approach to traffic prediction using multi-objective federated learning. Instead of relying on a centralized cloud server for data processing, collaborative training is implemented among several participants. The proposed method utilizes the advantages of reinforcement learning in dynamic decision-making scenarios and the expressive capabilities of graphical models to identify traffic intensity. Furthermore, a new methodology integrates federated learning concepts with multi-objective optimization to forecast traffic patterns accurately. The proposed approach exhibits a higher level of performance than existing methods for estimating traffic speed. It achieves a communication delay of 23.4%, packet delivery ratio (PDR) of 92.45%, packet loss rate of 12.34%, prediction accuracy of 97.45%, and resource utilization of 89.56%. The visualisation findings demonstrate that this new approach is able to successfully capture interconnections of metropolitan areas in different neighboring cities.