The distributional learning theory or learning of probability distribution is a framework in computational learning theory. It has been proposed from Michael Kearns, , Dana Ron, Ronitt Rubinfeld, Robert Schapire and in 1994 and it was inspired from the PAC-framework introduced by Leslie Valiant. In this framework the input is a number of samples drawn from a distribution that belongs to a specific class of distributions. The goal is to find an efficient algorithm that, based on these samples, determines with high probability the distribution from which the samples have been drawn. Because of its generality, this framework has been used in a large variety of different fields like machine learning, approximation algorithms, applied probability and statistics. This article explains the basic definitions, tools and results in this framework from the theory of computation point of view. (Wikipedia).
(PP 6.1) Multivariate Gaussian - definition
Introduction to the multivariate Gaussian (or multivariate Normal) distribution.
From playlist Probability Theory
The Complexity of Distributions - Emanuele Viola
Emanuele Viola Northeastern University March 5, 2012 Complexity theory, with some notable exceptions, typically studies the complexity of computing a function h(x) of a *given* input x. We advocate the study of the complexity of generating -- or sampling -- the output distribution h(x) for
From playlist Mathematics
(ML 7.7.A1) Dirichlet distribution
Definition of the Dirichlet distribution, what it looks like, intuition for what the parameters control, and some statistics: mean, mode, and variance.
From playlist Machine Learning
Distributions - Statistical Inference
In this video I talk about distribution, how to visualize it and also provide a concrete definition for it.
From playlist Statistical Inference
(ML 3.7) The Big Picture (part 3)
How the core concepts and methods in machine learning arise naturally in the course of solving the decision theory problem. A playlist of these Machine Learning videos is available here: http://www.youtube.com/my_playlists?p=D0F06AA0D2E8FFBA
From playlist Machine Learning
Breaking the Communication-Privacy-Accuracy Trilemma
A Google TechTalk, 2020/7/29, presented by Ayfer Ozgur Aydin, Stanford University ABSTRACT: Two major challenges in distributed learning and estimation are 1) preserving the privacy of the local samples; and 2) communicating them efficiently to a central server, while achieving high accura
From playlist 2020 Google Workshop on Federated Learning and Analytics
What is a Sampling Distribution?
Intro to sampling distributions. What is a sampling distribution? What is the mean of the sampling distribution of the mean? Check out my e-book, Sampling in Statistics, which covers everything you need to know to find samples with more than 20 different techniques: https://prof-essa.creat
From playlist Probability Distributions
(PP 6.6) Geometric intuition for the multivariate Gaussian (part 1)
How to visualize the effect of the eigenvalues (scaling), eigenvectors (rotation), and mean vector (shift) on the density of a multivariate Gaussian.
From playlist Probability Theory
Understanding the Central Limit Theorem
From playlist Unit 7 Probability C: Sampling Distributions & Simulation
Benjamin Guedj: On generalisation and learning
A (condensed) primer on PAC-Bayes, followed by News from the PAC-Bayes frontline. LMS Computer Science Colloquium 2021
From playlist LMS Computer Science Colloquium Nov 2021
Probability & Information Theory — Subject 5 of Machine Learning Foundations
#MLFoundations #Probability #MachineLearning Welcome to my course on Probability and Information Theory, which is part of my broader "Machine Learning Foundations" curriculum. This video is an orientation to the curriculum. There are eight subjects covered comprehensively in the ML Found
From playlist Probability for Machine Learning
Statistical mechanics of deep learning by Surya Ganguli
Statistical Physics Methods in Machine Learning DATE: 26 December 2017 to 30 December 2017 VENUE: Ramanujan Lecture Hall, ICTS, Bengaluru The theme of this Discussion Meeting is the analysis of distributed/networked algorithms in machine learning and theoretical computer science in the
From playlist Statistical Physics Methods in Machine Learning
Statistical Learning Theory for Modern Machine Learning - John Shawe-Taylor
Seminar on Theoretical Machine Learning Topic: Statistical Learning Theory for Modern Machine Learning Speaker: John Shawe-Taylor Affiliation: University College London Date: August 11, 2020 For more video please visit http://video.ias.edu
From playlist Mathematics
Learning probability distributions; What can, What can't be done - Shai Ben-David
Seminar on Theoretical Machine Learning Topic: Learning probability distributions; What can, What can't be done Speaker: Shai Ben-David Affiliation: University of Waterloo Date: May 7, 2020 For more video please visit http://video.ias.edu
From playlist Mathematics
A New Physics-Inspired Theory of Deep Learning | Optimal initialization of Neural Nets
A special video about recent exciting developments in mathematical deep learning! 🔥 Make sure to check out the video if you want a quick visual summary over contents of the “The principles of deep learning theory” book https://deeplearningtheory.com/. SPONSOR: Aleph Alpha 👉 https://app.al
From playlist Explained AI/ML in your Coffee Break
Phiala Shanahan: "Machine learning for lattice field theory"
Machine Learning for Physics and the Physics of Learning 2019 Workshop I: From Passive to Active: Generative and Reinforcement Learning with Physics "Machine learning for lattice field theory" Phiala Shanahan, Massachusetts Institute of Technology (MIT) Abstract: I will discuss opportuni
From playlist Machine Learning for Physics and the Physics of Learning 2019
Seminar 9: Surya Ganguli - Statistical Physics of Deep Learning
MIT RES.9-003 Brains, Minds and Machines Summer Course, Summer 2015 View the complete course: https://ocw.mit.edu/RES-9-003SU15 Instructor: Surya Ganguli Describes how the application of methods from statistical physics to the analysis of high-dimensional data can provide theoretical insi
From playlist MIT RES.9-003 Brains, Minds and Machines Summer Course, Summer 2015
Tightening information-theoretic generalization bounds with data-dependent estimate... - Daniel Roy
Workshop on Theory of Deep Learning: Where next? Topic: Tightening information-theoretic generalization bounds with data-dependent estimates with an application to SGLD Speaker: Daniel Roy Affiliation: University of Toronto Date: October 15, 2019 For more video please visit http://video
From playlist Mathematics
ICML 2018: Tutorial Session: Toward the Theoretical Understanding of Deep Learning
Watch this video with AI-generated Table of Content (ToC), Phrase Cloud and In-video Search here: https://videos.videoken.com/index.php/videos/icml-2018-tutorial-session-toward-the-theoretical-understanding-of-deep-learning/
From playlist ML @ Scale
(PP 6.7) Geometric intuition for the multivariate Gaussian (part 2)
How to visualize the effect of the eigenvalues (scaling), eigenvectors (rotation), and mean vector (shift) on the density of a multivariate Gaussian.
From playlist Probability Theory