A5020252920- Recommender Systems and Techniques
- Advanced Graph Neural Networks
- Topic Modeling
- Advanced Bandit Algorithms Research
- Expert finding and Q&A systems
- Diverse Legal and Medical Studies
- Generative Adversarial Networks and Image Synthesis
- Image Retrieval and Classification Techniques
- Data Quality and Management
- Mobile Crowdsensing and Crowdsourcing
- Natural Language Processing Techniques
- Smart Grid Energy Management
- Digitalization, Law, and Regulation
- Law and Political Science
University College Dublin
2021-2023
Politecnico di Milano
2019
In this paper we provide an overview of the approach used as team PoliCloud8 for ACM RecSys Challenge 2019. The competition, organized by Trivago, focuses on problem session-based and context-aware accommodation recommendation in a travel domain. goal is to suggest suitable accommodations fitting needs traveller maximise chance redirect (click-out) booking site, relying explicit implicit user signals within session (clicks, search refinement, filter usage) detect users intent. Our proposes...
Fashion industry is driven by fashion cycles, in which a item launched, rises to mainstream appeal and becomes trend, then diminishes eventually obsolete. These properties make it critical incorporate temporal information when adapting recommendation framework be employed the domain. However, an standard real-world architecture entails numerous phases, including data preparation, establishing training recommender models, filtering fulfilling revenue-based user needs. The contributions of...
Most state-of-the-art top-N collaborative recommender systems work by learning embeddings to jointly represent users and items. Learned are considered be effective solve a variety of tasks. Among others, providing explaining recommendations. In this paper we question the reliability learned Matrix Factorization (MF). We empirically demonstrate that, simply changing initial values assigned latent factors, same MF method generates very different items users, highlight that effect is stronger...
Rs4rs is a web application designed to perform semantic search on recent papers from top conferences and journals related Recommender Systems. Current scholarly engine tools like Google Scholar, Semantic ResearchGate often yield broad results that fail target the most relevant high-quality publications. Moreover, manually visiting individual conference journal websites time-consuming process primarily supports only syntactic searches. addresses these issues by providing user-friendly...
Recommender systems research lacks standardized benchmarks for reproducibility and algorithm comparisons. We introduce RBoard, a novel framework addressing these challenges by providing comprehensive platform benchmarking diverse recommendation tasks, including CTR prediction, Top-N recommendation, others. RBoard's primary objective is to enable fully reproducible reusable experiments across scenarios. The evaluates algorithms multiple datasets within each task, aggregating results holistic...
Rs4rs is a web application designed to perform semantic search on recent papers from top conferences and journals related Recommender Systems. Current scholarly engine tools like Google Scholar, Semantic ResearchGate often yield broad results that fail target the most relevant high-quality publications. Moreover, manually visiting individual conference journal websites time-consuming process primarily supports only syntactic searches. addresses these issues by providing user-friendly...
Reinforcement learning (RL) has demonstrated great potential to improve slate-based recommender systems by optimizing recommendations for long-term user engagement. To handle the combinatorial action space in slate recommendation, recent works decompose Q-value of a into item-wise Q-values, using an value-based policy. However, common case where value function is parameterized taking state and as input results linearly increasing number evaluations required select action, proportional...
Most state-of-the-art top-N collaborative recommender systems work by learning embeddings to jointly represent users and items. Learned are considered be effective solve a variety of tasks. Among others, providing explaining recommendations. In this paper we question the reliability learned Matrix Factorization (MF). We empirically demonstrate that, simply changing initial values assigned latent factors, same MF method generates very different items users, highlight that effect is stronger...