ACM Europe Council Best Paper Award to the Politecnico di Milano

An analysis of deep learning algorithms applied to recommender systems by Ferrari Dacrema and Cremonesi


Maurizio Ferrari Dacrema, an Information Technology PhD student, and Prof. Paolo Cremonesi, from the Architecture Research Team of the Electronics, Information, and Bioengineering Department, won the ACM Europe Council Best Paper Awardat the 13th ACM Recommender Systems conference.

The article "Are We Really Making Much Progress? A Worrying Analysis of Recent Neural Recommendation Approaches" provides a systematic analysis of the deep learning algorithms applied to recommender systems, published at high level conferences in recent years.

The article identifies 18 significant publications, stating that only 7 contain results that can be reproduced, and 6 of them were found to be non-competitive compared to much more specific non-neural techniques.

Overall, the article highlights various potential problems with the use of machine learning, and maintains that experimental practices need to be improved in this sector.

The ACM Recommender Systems (RecSys) conference is the leading international forum for presenting new research results, systems, and techniques in the recommender system field.

Recommender systems are information filtering systems intended for users, in order to facilitate browsing of vast catalogues of information, products and news, now available in the information and e-commerce era. Suggestions used by the user are personalised based on past interactions with the system, their preferences, and other personal information.

 

For more information:
The winning paper