News – CDL-RecSys

Choice-Based Preference Elicitation to Reduce the Cold Start Problem of a Leisure Activities Recommender in a Mobile App

Author: Andreas Fink Supervisor: Julia Neidhardt Abstract Early stages of user interactions pose a significant challenge for recommender systems, particularly due to the cold start problem. This issue arises when new users or items lack historical data, making it difficult to generate personalized recommendations. This thesis investigates the effectiveness of four visual preference elicitation methods…

Comparative Analysis of Fashion Captioning and Multimodal Fashion Recommendation

Author: Maria De Los Angeles Gwendolyn Aglae Rippberger Fonseca Supervisor: Julia Neidhardt Abstract This thesis explores two main tasks: (1) fine-tuning image captioning models for fashion datasets and (2) evaluating different feature spaces for personalized fashion recommendations. We fine-tune state-of-the-art vision-language models – BLIP-2 and LLaVA – on two fashion datasets, H&M and FACAD, to…

Context over Categories – Implementing the Theory of Constructed Emotion

Author: Nils Klüwer Supervisor: Julia Neidhardt, Co-Supervisor: Irina Nalis-Neuner Abstract Emotion analysis is a critical research area with applications ranging from content moderation to personalized systems. Despite its importance, many approaches rely on traditional models, such as Ekman’s universal emotions theory, which reduces emotions to static, predefined categories. This oversimplification neglects the complexity and contextual…