Kreikemeyer, Justin N. and Skrzypczak, Glenn and Tominski, Christian and Uhrmacher, Adelinde M. (2026) Interactive Visual Inference of Chemical Reaction Networks From Time-Series Data. In: 2026 Winter Simulation Conference, 6-9 Dec 2026, Glasgow, Scotland, UK.
Full text not available from this repository.Abstract
Recent advances in artificial intelligence automate many scientific tasks, including the inference of quantitative simulation models from time-series data. Although many inference methods are powerful, they are often opaque to practitioners and neither guarantee optimal nor unique reconstructions. Consequently, it makes sense to bring human judgment back into the loop. To this end, we describe a method for interactive visual inference based on the sparse regression principle. This also enables a playful exploration of challenges with inverse problems and the inductive modeling workflow by making the relationship between a model’s components and its overall behavior tangible. As target formalism, we employ chemical reaction networks where the components are reactions. Their unique coupling property allows informing the user about the influence of reactions without much experimentation. We implement our concept as the prototype ReactionQuest and first feedback shows positive effects on users’ understanding of model inference.
| Item Type: | Conference or Workshop Item (Paper) |
|---|---|
| Additional Information: | accepted |
