VAxAutoSci: Visual Analytics in the Age of Autonomous Science
Organizers:Shayan Monadjemi,Gabriel Appleby,Quan Nguyen,Ayana Ghosh,Christoph Heinzl,Remco Chang
Artificial intelligence is rapidly transforming scientific workflows. In emerging self-driving laboratories (SDLs), autonomous agents design experiments, analyze results, and iteratively refine hypotheses within closed-loop pipelines, fundamentally shifting the role of the scientist. This transition creates new opportunities for visual analytics to enable oversight and steering of autonomous processes, facilitate the inspection and refinement of machine-generated hypotheses, and support effective human–AI collaboration in scientific discovery. This workshop positions visual analytics as a core enabler of autonomous scientific discovery and advances two complementary directions: (1) developing methods that support AI-accelerated science, and (2) leveraging AI-accelerated scientific platforms to advance visualization research into AI-driven workflows. We will encourage submissions at the intersection of visual analytics, self-driving labs, and scientific domains (e.g., materials science). The workshop will include an invited keynote presentation, paper presentations, demos, and group discussions that help us articulate a concrete research agenda for visual analytics in the age of autonomous science.
Papers
Defaulting to Bar Charts: Chart-Type Bias and Scalar-Comparison Collapse in Autonomous Visualization Agents
MolExplain: An Interactive Tool for Explainable Molecular Property Prediction
Spatial Action Review: A Visual Analytics Dashboard for Auditing Language-to-Action Hand-offs in Electron Microscopy
Authors: Samia Mohinta, Albert Cardona
Toward a Scientific Discovery Engine for Weather and Climate Data: A Visual Analytics Workbench for Embedding-Based Exploration
ViSAXS: Visual Analytics for Human-Guided SAXS Morphology Exploration in Autonomous Characterization
Authors: Aishik Deb, Kevin Yager, Klaus Mueller