Single-Branch Network Architectures to Close the Modality Gap in Multimodal Recommendation
ACM Transactions on Recommender Systems
Student Research Assistant
Christian Ganhör, MSc
Johannes Kepler University Linz
About
During my studies at Johannes Kepler University Linz, I worked as a student research assistant at the Institute of Computational Perception. My research focused on Recommender Systems - initially on mitigating inherent biases, and later on leveraging multimodal side information to address cold-start and missing-modality scenarios. My work has been published in renowed conferences and journals, including SIGIR, RecSys, and TORS.
In addition, I contributed to EmoMTB, a science communication project presented at the Ars Electronica Festival, with which we showcased an interactive music exploration interface.
For the latest updates and my contact details, please visit my website.
Research output
Publications
A Multimodal Single-Branch Embedding Network for Recommendation in Cold-Start and Missing Modality Scenarios
Proceedings of the 18th ACM Conference on Recommender Systems (RecSys)
Simultaneous Unlearning of Multiple Protected User Attributes From Variational Autoencoder Recommenders Using Adversarial Training
Proceedings of the 5th International Workshop on Algorithmic Bias in Search and Recommendation (BIAS @ SIGIR 2024)
Emotion-aware Music Tower Blocks (EmoMTB): An Intelligent Audiovisual Interface for Music Discovery and Recommendation
International Journal of Multimedia Information Retrieval (IJMIR)
EmoMTB: Emotion-aware Music Tower Blocks
Proceedings of the 2022 ACM International Conference on Multimedia Retrieval (ICMR)
ProtoMF: Prototype-based Matrix Factorization for Effective and Explainable Recommendations
Proceedings of the 16th ACM Conference on Recommender Systems (RecSys)
Unlearning Protected User Attributes in Recommendations with Adversarial Training
Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval