Modern online services continuously generate data at very fast rates. This continuous flow of data encompasses content - e.g. posts, news, products, comments -, but also user feedback - e.g. ratings, views, reads, clicks -, together with context data. This can be overwhelming for systems and algorithms designed to train in batches, given the continuous and unpredictable rate of change of content, context and user preferences or intents, especially in long-term modeling. Therefore, it is important to investigate online methods able to transparently and robustly adapt to the multitude of dynamics of online services.


Incremental models and online learning methods are gaining attention in the recommender systems community, given their natural ability to deal with the continuous flows of data generated in dynamic, complex environments. The objective of this workshop is to foster contributions and bring together a growing community of researchers and practitioners interested in online, adaptive approaches to short- and long-term user modeling, recommendation and personalization, and their evaluation regarding multiple dimensions, such as fairness, privacy, explainability, and reproducibility.


Relevant topics include, but are not limited to:

- Stream-based and incremental algorithms
- Continual learning and forgetting
- Lifelong user modeling and recommendation
- User preference change detection and adaptation
- Context change detection and adaptation
- Session-based and sequential learning
- Online distributed and decentralized models
- Online learning with bandits and reinforcement learning
- Online learning from evolving graphs
- Online automated ML
- Online counterfactual learning
- Time-sensitive recommendation
- Privacy and user sovereignty in incremental models
- Interpretability of evolving models
- Online evaluation and benchmarking
- Bias evolution monitoring
- Reproducibility in online methods
- Scalability of online algorithms
- Platforms, software, data, and architectures
- Industrial case studies