The Eighth Data Science Meets Optimization (DSO) Workshop at IJCAI 2026 in Bremen, Germany

Data science and optimization are tightly intertwined. Many problems in data science can be solved using optimizers. On the other side, most optimization problems stated through classical models such as those from mathematical programming cannot be considered independent of historical data. Examples are ample: Methods aimed at high-level combinatorial optimization have been shown to strongly benefit from configuration, algorithm selection and tuning tools that are learned from historical data; Machine Learning (ML) often relies on optimization techniques such as linear or integer programming, and increasingly so for verification and optimal decision trees; Metaheuristic approaches characterized by learning components are commonplace in mathematical optimization; Black-box optimization makes heavy use of machine learning; deep learning is increasingly adopted to predict solutions to combinatorial problems (such as routing and scheduling problems). Furthermore, ML models are embedded into the combinatorial optimization pipeline to address hard-to-model systems, to validate the ML model itself, or intertwined in methodologies like decision-focused learning. Not least, LLMs are used to help formulate and solve optimization problems, and optimization techniques can help verify LLM outputs, for example.