Generalization Error and Stability
Explore statistical learning theory, focusing on generalization error, stability, and empirical risk minimization. Gain insights into universal consistency and the law of large numbers.
Understanding Deep Learning - Challenges for Statistical Learning Theory
Explore statistical learning theory challenges in deep learning, focusing on key concepts and recent developments in this complex field.
Learning from Dependent Data - Lecture 1
Explore techniques for learning from dependent data, focusing on probabilistic and optimization methods in data science. Gain insights into handling interconnected information effectively.
Quantum Statistical Query Learning II of II - IPAM at UCLA
Explores quantum statistical query learning, comparing it to quantum PAC models. Discusses lower bounds, separation results, and potential applications in quantum learning theory.
On the Variance and Admissibility of Empirical Risk Minimization on Convex Classes
Explores empirical risk minimization's variance and admissibility in function estimation, proving its minimax optimal variance and admissibility using concentration of measure techniques in convex geometry.
Robust Learning of a Single Neuron - Bridging Computational Gaps Using Optimization
Explore robust learning of a single neuron with ReLU activation, focusing on optimization techniques to achieve near-optimal mean square loss in adversarial settings. Discover novel approaches for various activation functions.
Testing Assumptions of Learning Algorithms
Explore how property testing can mitigate reliance on distributional assumptions and no distribution shift in supervised learning algorithms, ensuring more reliable theoretical guarantees.
Random Tessellation Forests: Overcoming the Curse of Dimensionality
Explore random tessellation forests in machine learning, overcoming dimensionality challenges through innovative partitioning methods and directional distributions for improved statistical performance.
Learning to Reconstruct - Statistical Learning Theory and Encrypted Database Attacks
Explores the connection between statistical learning theory and encrypted database attacks, presenting novel approaches for database reconstruction from access pattern leakage with improved efficiency and broader applicability.
Boosting and Ensembles in Machine Learning - Lecture 18b
Discover how to transform weak learning algorithms into powerful predictive models through boosting and ensemble techniques in machine learning.
Can Generalized Divergences Help for Invariant Neural Networks?
Explore the potential of generalised divergences in enhancing invariant neural networks, examining their impact on network performance and robustness.
Local Glivenko-Cantelli: Estimating the Mean in Infinite Dimensions
Delve into advanced statistical theory exploring mean estimation in infinite dimensions, focusing on Local Glivenko-Cantelli bounds and maximal deviation analysis for Boolean cube distributions.
Mean Estimation in High-Dimensional Binary Markov Gaussian Mixture Models
Delve into high-dimensional mean estimation in binary Markov Gaussian mixture models, exploring how memory in data affects estimation error and the interplay between sample size, dimension, and signal strength.
Supervised Learning: The Setup and Hypothesis Space - Lecture 3
Delve into supervised learning concepts, focusing on hypothesis space selection and its crucial role in machine learning model development and performance optimization.
Compatibility and the Lasso - Lecture 2
Explore compatibility and the Lasso in statistical learning with Sara van de Geer from ETH Zurich in this distinguished lecture on advanced machine learning techniques.
Overfitting: Benign, Tempered and Harmful - Lecture on Machine Learning Regularization
Explore the nuanced understanding of overfitting in neural networks, examining benign, tempered, and harmful effects on model performance based on data properties and training conditions.
Mathematical Problems in Machine Learning - Part 2 of 4
Explore mathematical foundations of machine learning with Andrea Montanari, focusing on key problems and theoretical concepts in the field.
Mathematical Problems in Machine Learning - Part 1 of 4
Explore mathematical foundations of machine learning with Andrea Montanari, focusing on key problems and theoretical concepts in the field.
Benign Overfitting in Linear Regression and Deep Learning
Explore benign overfitting in deep learning with Peter Bartlett. Uncover insights on linear regression, effective rank, and overparameterization in neural networks for optimal prediction accuracy.
Modern Paradigms of Generalization - The Heliocentric Model of Aristarchus
Explore modern generalization paradigms and historical scientific models in this insightful lecture on machine learning theory and its historical parallels.
Exploiting Sparsity and Structure in Parametric and Nonparametric Estimation - 2007
Explore advanced techniques in sparse estimation for graphical models and nonparametric regression, including l1 regularization and derivative regularization for high-dimensional data analysis.