MSML21

A Data Driven Method for Computing Quasipotentials

A 10-minute video from MSML21: “A Data Driven Method for Computing Quasipotentials”.

A Qualitative Study of the Dynamic Behavior for Adaptive Gradient Algorithms

A 13-minute video from MSML21: “A Qualitative Study of the Dynamic Behavior for Adaptive Gradient Algorithms”.

A deep learning method for solving Fokker-Planck equations

Advanced

A 14-minute video from MSML21: “A deep learning method for solving Fokker-Planck equations”.

A semigroup method for high dimensional committor functions based on neural network

Advanced

A 15-minute video from MSML21: “A semigroup method for high dimensional committor functions based on neural network”.

Active learning with importance sampling

A 17-minute video from MSML21: “Active learning with importance sampling”.

Adversarial Robustness of Stabilized Neural ODE Might be from Obfuscated Gradients

Advanced

A 14-minute video from MSML21: “Adversarial Robustness of Stabilized Neural ODE Might be from Obfuscated Gradients”.

Analyzing Finite Neural Networks: Can We Trust Neural Tangent Kernel Theory?

Advanced

A 15-minute video from MSML21: “Analyzing Finite Neural Networks: Can We Trust Neural Tangent Kernel Theory?”.

Average-Case Integrality Gap for Non-Negative Principal Component Analysis

A 15-minute video from MSML21: “Average-Case Integrality Gap for Non-Negative Principal Component Analysis”.

BEAR: Sketching BFGS Algorithm for Ultra-High Dimensional Feature Selection in Sublinear Memory

Advanced

A 12-minute video from MSML21: “BEAR: Sketching BFGS Algorithm for Ultra-High Dimensional Feature Selection in Sublinear Memory”.

Borrowing From the Future: Addressing Double Sampling in Model-free Control

A 15-minute video from MSML21: “Borrowing From the Future: Addressing Double Sampling in Model-free Control”.

Construction of optimal spectral methods in phase retrieval

Advanced

A 15-minute video from MSML21: “Construction of optimal spectral methods in phase retrieval”.

DNNs Are Effective At Learning High-Dimensional Hilbert-Valued Functions From Limited Data

A 15-minute video from MSML21: “DNNs Are Effective At Learning High-Dimensional Hilbert-Valued Functions From Limited Data”.

Decentralized Multi-Agents by Imitation of a Centralized Controller

A 13-minute video from MSML21: “Decentralized Multi-Agents by Imitation of a Centralized Controller”.

Deep Autoencoders: From Understanding to Generalization Guarantees

A 12-minute video from MSML21: “Deep Autoencoders: From Understanding to Generalization Guarantees”.

Deep Generative Learning via Euler Particle Transport

A 14-minute video from MSML21: “Deep Generative Learning via Euler Particle Transport”.

Dynamic Algorithms for Online Multiple Testing

A 13-minute video from MSML21: “Dynamic Algorithms for Online Multiple Testing”.

Generalization and Memorization: The Bias Potential Model

Advanced

A 14-minute video from MSML21: “Generalization and Memorization: The Bias Potential Model”.

Ground States of Quantum Many Body Lattice Models via Reinforcement Learning

Advanced

A 12-minute video from MSML21: “Ground States of Quantum Many Body Lattice Models via Reinforcement Learning”.

Hessian Estimation via Stein's Identity in Black-Box Problems

A 15-minute video from MSML21: “Hessian Estimation via Stein's Identity in Black-Box Problems”.

Hessian-Aided Random Perturbation (HARP) Using Noisy Zeroth-Order Queries

A 18-minute video from MSML21: “Hessian-Aided Random Perturbation (HARP) Using Noisy Zeroth-Order Queries”.

Implicit form neural network for learning scalar hyperbolic conservation laws

Advanced

A 13-minute video from MSML21: “Implicit form neural network for learning scalar hyperbolic conservation laws”.

Interpretable and Learnable Super-Resolution Time-Frequency Representation

A 15-minute video from MSML21: “Interpretable and Learnable Super-Resolution Time-Frequency Representation”.

Kernel-Based Smoothness Analysis of Residual Networks

Advanced

A 16-minute video from MSML21: “Kernel-Based Smoothness Analysis of Residual Networks”.

Multilevel Stein variational gradient descent with applications to Bayesian inverse problems

A 13-minute video from MSML21: “Multilevel Stein variational gradient descent with applications to Bayesian inverse problems”.

Noise-Robust End-to-End Quantum Control using Deep Autoregressive Policy Networks

Advanced

A 16-minute video from MSML21: “Noise-Robust End-to-End Quantum Control using Deep Autoregressive Policy Networks”.

Numerical Calabi-Yau metrics from holomorphic networks

Advanced

A 16-minute video from MSML21: “Numerical Calabi-Yau metrics from holomorphic networks”.

On the emergence of tetrahedral symmetry in the final and penultimate layers (...)

A 15-minute video from MSML21: “On the emergence of tetrahedral symmetry in the final and penultimate layers (...)”.

Optimal Policies for a Pandemic: A Stochastic Game Approach and a Deep Learning Algorithm

Advanced

A 15-minute video from MSML21: “Optimal Policies for a Pandemic: A Stochastic Game Approach and a Deep Learning Algorithm”.

Parameter Estimation with Dense and Convolutional Neural Networks Applied to the FitzHugh-Nagumo ODE

Advanced

A 17-minute video from MSML21: “Parameter Estimation with Dense and Convolutional Neural Networks Applied to the FitzHugh-Nagumo ODE”.

Phase Retrieval with Holography and Untrained Priors:

Advanced

A 15-minute video from MSML21: “Phase Retrieval with Holography and Untrained Priors:”.

Practical and Fast Momentum-Based Power Methods

Advanced

A 15-minute video from MSML21: “Practical and Fast Momentum-Based Power Methods”.

Reconstruction of Pairwise Interactions using Energy-Based Models

Advanced

A 15-minute video from MSML21: “Reconstruction of Pairwise Interactions using Energy-Based Models”.

Reduced Order Modeling using Shallow ReLU Networks with Grassmann Layers

Advanced

A 14-minute video from MSML21: “Reduced Order Modeling using Shallow ReLU Networks with Grassmann Layers”.

Sharp threshold for alignment of graph databases with Gaussian weights

Advanced

A 13-minute video from MSML21: “Sharp threshold for alignment of graph databases with Gaussian weights”.

Solvable Model for Inheriting the Regularization through Knowledge Distillation

Advanced

A 12-minute video from MSML21: “Solvable Model for Inheriting the Regularization through Knowledge Distillation”.

Solving Bayesian Inverse Problems via Variational Autoencoders

A 15-minute video from MSML21: “Solving Bayesian Inverse Problems via Variational Autoencoders”.