Institute for Advanced Study Courses

Sparse Matrices in Sparse Analysis - Anna Gilbert

Explore sparse matrices in compressive sensing, focusing on expander graphs and their applications. Learn about metric repair scenarios and algorithms for sparse signal recovery.

• YouTube
• 1 hour 2 minutes
• On-Demand
• Free Video

Primes and Knots - Akshay Venkatesh

Explore surprising parallels between number theory and 3D geometry, uncovering connections that continue to shape mathematical research.

• YouTube
• 51 minutes
• On-Demand
• Free Video

Optimization Landscape and Two-Layer Neural Networks - Rong Ge

Explore optimization landscapes and two-layer neural networks, focusing on non-convexity, saddle points, and symmetric distributions in theoretical machine learning.

• YouTube
• 58 minutes
• On-Demand
• Free Video

Red Supergiants - New Perspectives on Dying Stars - Emily Levesque

Explore new perspectives on red supergiants, their characteristics, and significance in stellar evolution. Learn about observational techniques, binary interactions, and implications for supernova research.

• YouTube
• 57 minutes
• On-Demand
• Free Video

Toward a Causal Analysis of Generalization in Deep Learning - Behnam Neyshabur

Explore causal factors behind deep learning generalization, examining measures like path norm, flatness, and gradient noise to predict and understand model performance.

• YouTube
• 33 minutes
• On-Demand
• Free Video

Representational Power of Graph Neural Networks - Stefanie Jegelka

Explore the representational power of Graph Neural Networks, focusing on message passing, aggregation operations, and their impact on network structure and reasoning tasks.

• YouTube
• 28 minutes
• On-Demand
• Free Video

Provably Efficient Reinforcement Learning with Linear Function Approximation - Chi Jin

Explore efficient reinforcement learning with linear function approximation, covering key concepts like sample efficiency, value-based algorithms, and exploration in multi-armed bandits and linear MDPs.

• YouTube
• 28 minutes
• On-Demand
• Free Video

Learning Representations Using Causal Invariance - Leon Bottou

Explore causal invariance in machine learning representations, focusing on stable properties, extrapolation, and robust supervised learning techniques for improved model performance and generalization.

• YouTube
• 33 minutes
• On-Demand
• Free Video

From Classical Statistics to Modern ML - The Lessons of Deep Learning - Mikhail Belkin

Explore the transition from classical statistics to modern machine learning, focusing on deep learning's impact, interpolation methods, and the "double descent" phenomenon in generalization theory.

• YouTube
• 37 minutes
• On-Demand
• Free Video

On the Connection Between Neural Networks and Kernels: A Modern Perspective - Simon Du

Explore connections between neural networks and kernels, focusing on over-parameterization, generalization, and neural tangent kernels. Insights on training dynamics and applications to various learning tasks.

• YouTube
• 31 minutes
• On-Demand
• Free Video

Energy-Based Approaches to Representation Learning - Yann LeCun

Explore energy-based approaches to representation learning with Yann LeCun, covering self-supervised learning, video prediction, and sparse modeling in high-dimensional spaces.

• YouTube
• 40 minutes
• On-Demand
• Free Video

Are All Features Created Equal? - Aleksander Madry

Explore machine learning's success, deep learning's appeal, and adversarial perturbations. Examine robust features, human vs. ML perspectives, and implications for data efficiency and image synthesis.

• YouTube
• 37 minutes
• On-Demand
• Free Video

Overcoming the Curse of Dimensionality and Mode Collapse - Ke Li

Explore techniques to overcome dimensionality challenges and mode collapse in deep learning, focusing on GANs, maximum likelihood estimation, and efficient algorithms for high-dimensional spaces.

• YouTube
• 33 minutes
• On-Demand
• Free Video

PAC-Bayesian Approaches to Understanding Generalization in Deep Learning - Gintare Dziugaite

Explore PAC-Bayesian approaches for understanding generalization in deep learning, covering risk bounds, optimal priors, and data-dependent techniques for neural networks.

• YouTube
• 32 minutes
• On-Demand
• Free Video

Is Optimization the Right Language to Understand Deep Learning? - Sanjeev Arora

Explore deep learning optimization, generalization, and neural tangent kernels with Sanjeev Arora. Gain insights into training wide networks and matrix completion techniques.

• YouTube
• 33 minutes
• On-Demand
• Free Video