Advancing Science through AI

This workshop aims to catalyze progress at the intersection of artificial intelligence and the natural sciences by highlighting the methodological challenges and opportunities that arise when applying AI techniques to domain-specific problems in life, material, physical sciences, and related fields. Despite recent successes, the integration of AI into scientific workflows remains non-trivial due to domain constraints such as limited labeled data, complex simulation-based environments, and the need for interpretability and physical consistency.

The workshop will focus on several key objectives:

• Identify barriers to effective adoption of AI in scientific domains, including challenges in data representation, model generalization, integration with simulations, and alignment with domain knowledge.
• Showcase state-of-the-art approaches in AI-for-science, such as foundation models trained on multimodal scientific data, surrogate models for accelerating numerical simulations, AI-driven inverse design and optimization, integration of AI inference in-the-loop with simulation, and data-driven discovery in experimental settings.
• Foster exchange between AI researchers and domain scientists to align technical advances in machine learning with real-world scientific needs.
• Map future directions, emphasizing scalable, interpretable, and robust AI methods that are sensitive to the unique characteristics of scientific problems.