Recommender systems
- Summary
- Software that estimates, models and evaluates user preferences and needs, producing viable recommendations of items such as multimedia or goods that would match user interests.
Recommender systems are used in various user-facing services that involve diverse choices, such as the homepage of Netflix, the endless news feed at the bottom of the Seznam homepage, or “more like this” sections on Amazon. Our research in focuses on how to make recommendations more precise and computed more efficiently, as well as non-technical aspects, such as how users and stakeholders perceive the systems, and how the associated risks can be mitigated.
People
Patrik Dokoupil
Researcher
Ladislav Peška
Associate professor
Martin Spišák
Assistant researcher
Vojtěch Vančura
Researcher
Latest publications
- Efficient Learning of Sparse Representations from Interactions (2026)
- From Knots to Knobs: Towards Steerable Collaborative Filtering Using Sparse Autoencoders (2026)
- Towards Asynchronous Group Recommendations (2026)
- Match-it: Group Recommendations in the Age of Mobile Applications (2026)
- SM-RS 2.0: User-perceived Qualities of Single-and Multi-Objective Recommender Systems (2026)
- Long-term fairness in sequential group recommendations (2026)
- CALISSTA: Calibrated Candidate Retrieval Using Sparse Autoencoders and Top-k Aggregation (2026)
- GMAP 2026: 5th Workshop on Group Modeling, Adaptation and Personalization (2026)
- VISAnt: Unsupervised Data Exploration with Chernoff Faces (2025)
- It's Time to Let Go: Stopping Criteria Recommendations in Content-rich Domains (2025)
- Bridging the Rift: A Critical Perspective on the Divergence in Group Recommender Systems Research (2025)
- SAGEA: Sparse Autoencoder-based Group Embeddings Aggregation for Fairness-Preserving Group Recommendations (2025)
- How Do Users Perceive Recommender Systems’ Objectives? (2025)
- Effects of quantitative explanations on fairness perception in group recommender systems (2025)
- Accuracy and beyond-accuracy perspectives of controllable multi-objective recommender systems (2025)
- GMAP 2025: 4th Workshop on Group Modeling, Adaptation and Personalization (2025)
- On interpretability of linear autoencoders (2024)
- User perceptions of diversity in recommender systems (2024)
- SM-RS: Single-and multi-objective recommendations with contextual impressions and beyond-accuracy propensity scores (2024)
- Comparing user interfaces for customizing multi-objective recommender systems (2024)
Supporting software
Compresso
Compresso is an open-source PyTorch framework for sparse representation learning. It provides reusable building blocks for learning sparse neural representations, dynamic sparsification, sparse inference, and semantic analysis, enabling researchers to rapidly prototype sparse neural architectures while focusing on models rather than infrastructure.
EasyStudy
EasyStudy is an open-source framework for interactive user studies, including recommender systems, HCI, and generally any item-based experiments. The software includes many interesting algorithms and testing datasets.
GroupRec
GroupRec is a unified toolkit for reproducible and inspectable group recommendation research.