ReMAP Lab @ PoliMi is a research group at Politecnico di Milano focused on recommender systems. The group is part of the RecSys community, an international forum that meets annually at the RecSys conference.

ReMAP stands for Recommendation Models for Algorithmic Personalization. The name also nods to a few ideas: two standard evaluation metrics, Recall and Mean Average Precision, highlighting our focus on evaluation; the first step of many pipelines where we must ‘remap’ item or user identifiers to numerical ones; and, more broadly, a shift in perspective on how we approach recommendation problems.

Our current lines of research focus on recommender systems: evaluation and reproducibility, off-policy estimation for learning and model selection, and the integration of Large Language Models into recommendation pipelines. See the People page for more information. We also have a successful tradition of participating in the ACM RecSys Challenge every year since 2016, often with successful student teams, see our ACM RecSys Challenge page. 

NEWS: Politecnico di Milano’s team, coordinated by Maurizio Ferrari Dacrema and Andrea Pisani of our research group, won the first place of the academic part of the RecSys Challenge 2025.

NEWS: Amazon Personalize, a machine learning service that provides recommendation models, has added Hierarchical Recurrent Neural Network (HRNN), which our research group contributed to develop. For more details see the article and Amazon HRNN documentation.