Thema5127
Abschlussarbeitstyp: Bachelor, Master
Betreuer: Zhan Qu
Forschungsgruppe: Web Science
Archivierungsnummer: 5127
Abschlussarbeitsstatus: Offen
Beginn:
16. Januar 2024
Abgabe: unbekannt
In recent years, recommendation systems have become indispensable in our daily lives, facilitating users in discovering new content and products across diverse domains. How- ever, there exists a growing demand for recommendation systems that not only deliver accurate suggestions but also elucidate the rationale behind their recommendations. Moreover, while many recommendation systems concentrate on static interactions, real- world user engagements dynamically evolve over time.
This thesis endeavors to address these challenges by proposing an innovative counter- factual explanation method tailored for time-aware Graph Neural Network (GNN)-based recommender systems. Engaging in this research will provide an opportunity to gain profound insights into leveraging cutting-edge explainable Artificial Intelligence (AI) tech- niques for interpreting temporal graph neural networks. Furthermore, it offers a chance to comprehend the operational intricacies of recommender systems.
[1] https://www.mdpi.com/2076-3417/13/20/11176
Ausschreibung: Download (pdf)