ENABLING CROSS-DOMAIN TRANSFER OF USER PREFERENCES IN ELLIOT

This thesis explores the challenges of building recommendation models that can generalize across different domains. It presents an implementation of a cross-domain recommendation system using the Information Bottleneck Regularizer and Contrastive Information Regularizer implemented in the Elliot framework. We discuss the need for effective cross-domain recommendation systems and the challenges of building such systems. We introduce the Elliot framework, a versatile tool for building and evaluating recommendation systems, and explain its suitability for implementing our proposed model. We then describe the implementation of the cross-domain recommendation system using the Information Bottleneck Regularizer and Contrastive Information Regularizer and discuss the benefits of this technique. Finally, we conclude by discussing the potential applications of our implementation and future research directions in cross-domain recommendation systems.