Research Statement
My research investigates machine learning models for sequential and session-based recommendation from a graph-based perspective. Modern recommender systems must capture intricate temporal dynamics, short-term user intent, and long-term behavioral patterns. By formulating item interactions as graphs and hypergraphs, we can effectively discover multi-hop dependencies, repeated intent, and complex transition structures across user sessions.
In parallel, I explore Music Information Retrieval (MIR) and emotion recognition, applying multi-relational graph neural networks to model nuanced affective qualities from subjective listener annotations, tag structures, and audio representations.
Research Areas
Sequential Recommendation
Graph-based modeling of repeated intent and temporal user interactions.
Music Information Retrieval
Nuanced emotion recognition and music recommendation via graph neural networks.
Graph Neural Networks
Efficient shortest-path search, contrastive learning, and hypergraph representations.
Key Research Highlights & Projects
Efficient Session-based Recommendation with Contrastive Graph-based Shortest Path Search
Shortest-path global item relations and contrastive learning for scalable session-based recommendation. (ACM Transactions on Recommender Systems, 2025)
Nuanced Music Emotion Recognition via a Semi-Supervised Multi-Relational Graph Neural Network
Semi-supervised multi-relational graph neural networks to capture subtle emotion spaces from quality annotations. (Transactions of the International Society for Music Information Retrieval (8), 2025)
Hypergraph-based Temporal Modelling of Repeated Intent for Sequential Recommendation
Modeling repeated user intent and higher-order relations for sequential recommendation. (Proceedings of the ACM on Web Conference 2025, 2025)
Are We Explaining the Same Recommenders? Incorporating Recommender Performance for Evaluating Explainers
Investigating whether recommender performance shifts explainability metrics. (Proceedings of the 18th ACM Conference on Recommender Systems, 2024)
Venues
My research is published in venues including WWW, ACM RecSys, ACM TORS, UMAP, and TISMIR. It forms the scientific foundation behind my work as an AI engineer.