Lecture 2: Concentration of Measure
نویسنده
چکیده
In the proof, we used both finite-ness of H and the realizability assumptions in a crucial way. But we already saw one problem where the theorem does not apply, namely learning linear separators, where |H| =∞. Of course it is more common than not that realizability is not satisfied, so we’d also like some sort of guarantee in the absence of realizability. In the next few lectures we will build up the tools to relax both the realizability and the finite-ness assumptions. This requires more powerful probabilistic tools, in the form of concentration inequalities, and also takes us into the field of empirical process theory.
منابع مشابه
Concentration - of - measure inequalities Lecture notes by Gábor Lugosi
This text contains some of the material presented at the Summer
متن کاملNURSING CODES OF ETHICS EDUCATION BASED ON LECTURE ON MORAL PERFORMANCE OF NURSING STUDENTS: A QUASI-EXPERIMENTAL STUDY
Background & Aims: Nursing students as future nurses, are the largest provider of health services and will have a critical role in health care system. Therefore, with regard to the professionchr('39')s importance, Nurses along with clinical competencies, should have ethical competencies to provide appropriate and professional care. Therefore, the present study was conducted to determine the eff...
متن کاملStatistical Mechanics and Random Matrices
Statistical Mechanics and Random Matrices 3 1. Introduction 6 2. Motivations 7 3. The different scales; typical results 12 Lecture 1. Wigner matrices and moments estimates 15 1. Wigner's theorem 16 2. Words in several independent Wigner matrices 23 3. Estimates on the largest eigenvalue of Wigner matrices 25 Lecture 2. Gaussian Wigner matrices and Fredholm determinants 27 1. Joint law of the ei...
متن کاملCSE 291 : Learning Theory Fall 2006 Lecture 1 — Measure concentration
1.1 Concentration of measure: examples We start with some examples of “concentration of measure”. This phenomenon is very useful in analyzing machine learning algorithms and can be used to bound things like error probabilities. The first and the most standard of concentration results is for averages. It states that “the average of bounded independent random variables is tightly concentrated aro...
متن کاملSino-French Summer Institute: Stochastic Modeling and Applications
of Mini-course On the concentration properties of mean field particle models (based on a series of 4 joint works with : D.A. Dawson, A. Guionnet, E. Rio, S.L. Hu and L.M. Wu) Pierre Del Moral Centre de Recherche Bordeaux Sud-Ouest & Institut de Mathématique de Bordeaux INRIA, France Abstract This lecture is concerned with the exponential concentration properties of a general class of mean field...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2017