Institute for Pure & Applied Mathematics (IPAM) Courses

The Liquid Tensor Experiment - IPAM at UCLA

Explore the Liquid Tensor Experiment and its implications for machine-assisted proofs in mathematics, featuring innovative concepts and real-time mathematical demonstrations.

• YouTube
• 55 minutes
• Self-Paced
• Free Video

Bits and Atoms - Exploring the Intersection of Machine Learning and Microscopy

Explore the intersection of machine learning and microscopy, focusing on AI models for active learning in materials discovery, automated experiments, and physics discovery through advanced microscopy techniques.

• YouTube
• 48 minutes
• Self-Paced
• Free Video

Learning How to Break Symmetry With Symmetry-Preserving Neural Networks - IPAM at UCLA

Explore symmetry-preserving neural networks for data efficiency and generalization, learning symmetry-breaking information, and predicting structural distortions in crystalline materials.

• YouTube
• 1 hour
• Self-Paced
• Free Video

Diffusion Based Distributional Modeling of Conformers, Blind Docking and Proteins

Explore diffusion-based modeling for conformers, blind docking, and proteins. Learn about torsional diffusion, generative docking, and 3D motif scaffolding in molecular systems.

• YouTube
• 54 minutes
• Self-Paced
• Free Video

End-to-End Learning and Auto-Differentiation - Forces, Uncertainties, Etc.

Explore deep learning and auto-differentiation in molecular simulations, covering surrogate functions, sampling distributions, and time-propagation of differential equations for scientific problem-solving.

• YouTube
• 56 minutes
• Self-Paced
• Free Video

OpenFold - Lessons and Insights From Rebuilding and Retraining AlphaFold2

Insights from rebuilding AlphaFold2: OpenFold's optimized implementation, training process analysis, and implications for protein structure prediction and broader applications in structural biology.

• YouTube
• 48 minutes
• Self-Paced
• Free Video

Symbolic Regression for Discovery of a DFT Functional - IPAM at UCLA

Explore symbolic regression for DFT functional discovery, combining AutoML ideas to improve existing models in computational chemistry and materials science.

• YouTube
• 53 minutes
• Self-Paced
• Free Video

Advancing Molecular Simulation with Deep Learning - IPAM at UCLA

Explore cutting-edge applications of deep learning in molecular simulation, covering quantum chemistry, force fields, and Markov State Models for advanced scientific research.

• YouTube
• 59 minutes
• Self-Paced
• Free Video

AI in Cancer Research and Diagnostics - IPAM at UCLA

Explore AI's impact on cancer diagnostics, focusing on pathology applications, explainable methods, and multi-modal data integration for research and clinical use.

• YouTube
• 33 minutes
• Self-Paced
• Free Video

Interpretable Machine Learning via Program Synthesis - IPAM at UCLA

Explore interpretable machine learning through program synthesis, focusing on control policies and RNA splice prediction. Learn novel approaches for creating custom, problem-specific model families.

• YouTube
• 47 minutes
• Self-Paced
• Free Video

Machine Learning Enhanced Compressive Hyperspectral Imaging - IPAM at UCLA

Explore machine learning-enhanced compressive hyperspectral imaging techniques, combining optical systems and neural networks for efficient data acquisition and processing in various applications.

• YouTube
• 31 minutes
• Self-Paced
• Free Video

STFT Phase Retrieval - Robustness and Generative Priors - IPAM at UCLA

Explore STFT phase retrieval, focusing on robustness and generative priors. Learn about stable reconstruction techniques, application of generative models, and reducing required measurements in imaging and acoustics applications.

• YouTube
• 47 minutes
• Self-Paced
• Free Video

The Role of Data and Models for Deep-Learning Based Image Reconstruction

Explores deep learning in image reconstruction, focusing on model scaling, data requirements, and robustness. Discusses performance under distribution shifts and strategies for improvement.

• YouTube
• 46 minutes
• Self-Paced
• Free Video

Splitting the Difference Between Deep and Shallow Solutions of Inverse Problems

Explores balancing compressed sensing and deep learning in computational imaging, showcasing Plenoxels for efficient photorealistic view synthesis without neural networks.

• YouTube
• 46 minutes
• Self-Paced
• Free Video

Mesoscale Reconstruction of Images and Networks Using Tensor Decomposition

Explore mesoscale reconstruction of images and networks using tensor decomposition, focusing on a unified framework, error bounds, and online CP-dictionary learning for multi-modal datasets.

• YouTube
• 50 minutes
• Self-Paced
• Free Video