The volume contains the papers presented at the 15th Italian Information Retrieval Workshop (IIR 2025), held in Cagliari, Italy, from 3rd to 5th September 2025. This workshop continues the tradition of serving as a forum for researchers, academics, and industry professionals interested in the field of Information Retrieval (IR). The primary objective of IIR is to provide a platform for stimulating and disseminating research in IR, especially encouraging young researchers to present and discuss their work in an informal environment.
The workshop’s program featured 1 full paper, 2 short papers, and 24 extended abstracts. The authors of two contributions decided not to have their papers included in the proceedings. The event spanned three days and the presentations were organized into six thematic sessions: • Session 1 – Large Language Models and IR
System. LGOBE https://www.ludovicoboratto.com/ (L. Boratto); https://www.unibo.it/sitoweb/allegra.defilippo/ (A. De Filippo);
CEUR
ceur-ws.org • Session 2 – Recommender Systems: Models and Evaluation • Session 3 – Conversational and Personalized Systems – Federico Siciliano, Antonio Purificato, Filippo Betello, Nicola Tonellotto, Fabrizio Silvestri.
Caser+ and CosRec+: Closing the Gap Between CNNs and Attention Models in SRS. – Michael Benigni, Maurizio Ferrari Dacrema, Dietmar Jannach. Difusion Models for Recommendation: Reproducibility and Conceptual Mismatch. – Federico Siciliano, Giulia Di Teodoro, Nicola Tonellotto, Fabrizio Silvestri. Comparing
Recommendation Losses under Negative Sampling. – Neda Afreen, Ludovico Boratto, Gianni Fenu, Mirko Marras, Alessandro Soccol. Efective and Transparent Course Recommendation through Causal Reasoning with Language Models. – Alessandro Francesco Maria Martina, Alessandro Petruzzelli, Cataldo Musto, Marco de Gemmis, Pasquale Lops, Giovanni Semeraro. Beyond Homogeneous Users: Simulating Diverse User Personas for Conversational Recommendation Datasets. – Pranav Kasela, Gabriella Pasi, Rafaele Perego. PARK: Personalized Academic Retrieval with
Knowledge-graphs. – Marco Alessio, Simone Merlo, Tommaso Di Noia, Guglielmo Faggioli, Marco Ferrante, Nicola Ferro, Cristina Ioana Muntean, Franco Maria Nardini, Fedelucio Narducci, Rafaele Perego, Giuseppe Santucci, Nicola Viterbo. A Dataset for Joint Conversational Search and Recommendation. – Cristina Ioana Muntean, Franco Maria Nardini, Rafaele Perego, Guido Rocchietti, Cosimo
Rulli. Eficient Conversational Search via Topical Locality in Dense Retrieval . • Session 4 – Neural IR, Federated Models, and Search Applications – Georgeta Bordea, Stéphane Campinas, Matteo Catena, Renaud Delbru. Siren Federate:
Bridging Document, Relational, and Graph Models for Exploratory Graph Analysis. – Francesca Pezzuti, Sean MacAvaney, Nicola Tonellotto. Neural Web Crawling. – Marco Braga, Pranav Kasela, Alessandro Raganato, Gabriella Pasi. Investigating Task Arithmetic for Zero-Shot Information Retrieval. – Guglielmo Faggioli, Nicola Ferro, Rafaele Perego, Nicola Tonellotto. Counterfactual Dimension Importance Estimation (CoDIME) for Dense Information Retrieval. – Giulio D’Erasmo, Giovanni Trappolini, Nicola Tonellotto, Fabrizio Silvestri. Eclipse: Leveraging Pseudo-Irrelevance for Dimension Importance Estimation. • Session 5 – Evaluation, Metrics, and Tools – Francesco De Faveri, Guglielmo Faggioli, Nicola Ferro, Kalervo Järvelin. MDCU: A Multi
Dimensional Cumulative Utility Metric for Information Retrieval. – Silvio Martinico, Leonardo Delfino, Domenico Erriquez, Franco Maria Nardini, Rossano Venturini, Cosimo Rulli. kANNolo: Facilitating Research in Approximate Nearest Neighbors Search. – Andrea Pasin, Nicola Ferro. The KIMERA Infrastructure: Shifting from Evaluation-as-a-Service to Evaluation-in-the-Cloud. – Giuseppe Scarpi, Amir Khorrami Chokami, Diego Reforgiato. A Grading Methodology for
E-commerce Reviews based on a Mixed Lexical-semantic Approach. – Ornella Irrera, Stefano Marchesin, Gianmaria Silvello. Enhancing Ground Truth Creation for
IR Through Web-Based Collaborative Annotation. • Session 6 – User Modeling, Evaluation and Fairness
The program also included two invited talks: • “This Will All Make Sense When I’m Older”: Goals, Gaps, and the Road Ahead for Child-Aware IR
Maria Soledad (Sole) Pera – Delft University of Technology, Delft, The Netherlands. • Learned Sparse Representations of Text: A Story of Indexing and Retrieval
Franco Maria Nardini – Istituto di Scienza e Tecnologie dell’Informazione “Alessandro Faedo” del Consiglio Nazionale delle Ricerche (ISTI-CNR), Pisa, Italy.