Lecture theme: AI-for-Science: The next wave of artificial intelligence

Session 1:

Overview：AI-for-Science: The next wave of artificial

• AI Booming
• AI: From Mimicking Human to Discovering the World
• Science: Discover and Change the World
• Four Paradigms of Science
• Closing the Loop of Scientific Discovery

Session 2:

Solving Scientific Equations

• Graphormer for Molecular Modeling
• Molecular Simulation
• AI Models for Molecular Simulation
• Graphormer
• Graphormer: Results
• MD Simulation for SARS-COV-2
• The “Wedge” Effect of NTD
• DeepVortexNet for Fluid Modeling
• Meteorological Simulation
• Physics-informed Neural Network
• From Derivative-based to Monte-Carlo
• Deep Vortex Network

Session 3:

Mining Experimental Observations

• LorentzNet for Particle Detection
• Particle Detection
• Jet Analysis and Group Equivariance
• Problems with Existing AI Models
• LorentzNet
• Experimental Results
• SPT-Scientific Language Model
• SPT: Foundation Model for Science
• Effective Training: A System Work
• Effective Inference: Double Prompts
• Experimental Results

Session 4:

Discovering New Science

• AI for New Physics Detection
• Force Field Decomposing
• Learning Non-conservative Dynamics
• Experimental Results
• Target Applications
• More of Our Recent Research
• Microsoft Research AI4Science