5 papers
Training-Free, Identity-Preserving Image Editing for Fashion Pose Alignment and Normalization
Potito Aghilar, Vito Walter Anelli, Michelantonio Trizio +2
Diffusion models have recently unlocked new possibilities in editing images of real-world objects. Yet, transforming objects in non-rigid ways, such as modifying poses or applying…
Type-Less yet Type-Aware Inductive Link Prediction with Pretrained Language Models
Alessandro De Bellis, Salvatore Bufi, Giovanni Servedio +3
Inductive link prediction is emerging as a key paradigm for real-world knowledge graphs (KGs), where new entities frequently appear and models must generalize to them without retra…
XAI4LLM. Let Machine Learning Models and LLMs Collaborate for Enhanced In-Context Learning in Healthcare
Fatemeh Nazary, Yashar Deldjoo, Tommaso Di Noia +1
Clinical decision support systems require models that are not only highly accurate but also equitable and sensitive to the implications of missed diagnoses. In this study, we intro…
Do We Really Need Specialization? Evaluating Generalist Text Embeddings for Zero-Shot Recommendation and Search
Matteo Attimonelli, Alessandro De Bellis, Claudio Pomo +3
Pre-trained language models (PLMs) are widely used to derive semantic representations from item metadata in recommendation and search. In sequential recommendation, PLMs enhance ID…
Stealthy LLM-Driven Data Poisoning Attacks Against Embedding-Based Retrieval-Augmented Recommender Systems
Fatemeh Nazary, Yashar Deldjoo, Tommaso Di Noia +1
We present a systematic study of provider-side data poisoning in retrieval-augmented recommender systems (RAG-based). By modifying only a small fraction of tokens within item descr…