City-wide mobility prediction models typically rely on either training with extensive local trajectory data or applying transfer learning from data-rich cities to those with limited data. In both cases, the resulting models are specialized to a specific target city for which they require at least some trajectory data to adapt. Consequently, they cannot generalize to cities unseen during training and are inapplicable where no mobility data exist. In this work, we propose H 0 xtra, a novel approach that enables zero-shot trajectory prediction in entirely unseen cities. H0xtra leverages a hypernetwork to generate city-specific location embeddings from spatial distributions of points of interest, e.g., restaurants or stores. These embeddings capture city-agnostic location semantics, enabling a transformer to learn universal trajectory patterns across cities. At inference, H0xtra performs zeroshot transfer without requiring any mobility data or retraining. Adaptation requires only points of interest data, which are often publicly available, to generate location embeddings specific to the target city. Trained only on a small set of source cities, H0xtra achieves strong zero-shot generalization. In our experiments, the zero-shot performance achieves an accuracy improvement of 11.3% and an average displacement error reduction of 11.5% on average compared to state-of-the-art non-zero-shot baselines. Our code can be accessed at https://github.com/DLR-Prot-of-Terrestrial-Infrastructures/H0xtra.

User services increasingly base their actions on AI models, e.g., to offer personalized and proactive support. However, the underlying AI algorithms require a continuous stream of personal data—leading to privacy issues, as users typically have to share this data out of their territory. Current privacy-preserving concepts are either not applicable to such AI-based services or to the disadvantage of any party. This paper presents PrivAI, a new decentralized and privacy-by-design platform for overcoming the need for sharing user data to benefit from personalized AI services. In short, PrivAI complements existing approaches to personal data stores, but strictly enforces the confinement of raw user data. PrivAI further addresses the resulting challenges by (1) dividing AI algorithms into cloud-based general model training, subsequent local personalization, and community-based sharing of model updates for new users; by (2) loading confidential AI models into a trusted execution environment, and thus, protecting provider’s intellectual property (IP). Our experiments show the feasibility and effectiveness of PrivAI with comparable performance as currently-practiced approaches.

Many recommender systems use association rules mining, a technique that captures relations between user interests and recommends new probable ones accordingly. Applying association rule mining causes privacy concerns as user interests may contain sensitive personal information (e.g., political views). This potentially even inhibits the user from providing information in the first place. Current distributed privacy-preserving association rules mining (PPARM) approaches use cryptographic primitives that come with high computational and communication costs, rendering PPARM unsuitable for large-scale applications such as social networks. We propose improvements in the efficiency and privacy of PPARM approaches by minimizing the required data. We propose and compare sampling strategies to sample the data based on social graphs in a privacy-preserving manner. The results on real-world datasets show that our sampling-based approach can achieve a high average precision score with as low as 50% sampling rate and, therefore, with a 50% reduction of communication cost.

The importance of presentation recordings is steadily increasing. This trend is indicated for example by the growing MOOCs market. Many systems for the production of such recordings exist. However, produced recordings are not exchangeable between systems due to different representation formats. In this paper, we present an ontology for the conceptual description of presentation recordings and describe the transformation process between different systems. Furthermore, we explain how this ontology can be used to preserve presentation recordings as ebooks.