9.54 class 2. Shimon Ullman + Tomaso Poggio. Types of learning + biophysical mechanisms of plasticity. Danny Harari + Daniel Zysman + Darren Seibert

Similar documents
: Introduction to Machine Learning Dr. Rita Osadchy

Big Data Analytics for SCADA

Machine Learning. CUNY Graduate Center, Spring Professor Liang Huang.

Machine Learning: Overview

MA2823: Foundations of Machine Learning

Data, Measurements, Features

Software Development Training Camp 1 (0-3) Prerequisite : Program development skill enhancement camp, at least 48 person-hours.

CS 2750 Machine Learning. Lecture 1. Machine Learning. CS 2750 Machine Learning.

CS Master Level Courses and Areas COURSE DESCRIPTIONS. CSCI 521 Real-Time Systems. CSCI 522 High Performance Computing

BIOINF 585 Fall 2015 Machine Learning for Systems Biology & Clinical Informatics

Introduction to Machine Learning Lecture 1. Mehryar Mohri Courant Institute and Google Research

Machine Learning. Mausam (based on slides by Tom Mitchell, Oren Etzioni and Pedro Domingos)

Machine Learning Introduction

Master's projects at ITMO University. Daniil Chivilikhin PhD ITMO University

School of Computer Science

INTRODUCTION TO MACHINE LEARNING 3RD EDITION

Introduction to Machine Learning Using Python. Vikram Kamath

Maschinelles Lernen mit MATLAB

Machine Learning.

A Systemic Artificial Intelligence (AI) Approach to Difficult Text Analytics Tasks

An Introduction to Data Mining. Big Data World. Related Fields and Disciplines. What is Data Mining? 2/12/2015

Masters in Information Technology

Introduction to Pattern Recognition

Core Curriculum to the Course:

Assessing Data Mining: The State of the Practice

Network Machine Learning Research Group. Intended status: Informational October 19, 2015 Expires: April 21, 2016

Steven C.H. Hoi School of Information Systems Singapore Management University

6.2.8 Neural networks for data mining

School of Computer Science

Non-negative Matrix Factorization (NMF) in Semi-supervised Learning Reducing Dimension and Maintaining Meaning

Position Classification Flysheet for Computer Science Series, GS Table of Contents

4.1. Title: data analysis (systems analysis) Annotation of educational discipline: educational discipline includes in itself the mastery of the

BOOSTING - A METHOD FOR IMPROVING THE ACCURACY OF PREDICTIVE MODEL

ADVANCED MACHINE LEARNING. Introduction

Learning is a very general term denoting the way in which agents:

Master s Program in Information Systems

Course DSS. Business Intelligence: Data Warehousing, Data Acquisition, Data Mining, Business Analytics, and Visualization

Machine Learning and Statistics: What s the Connection?

TIETS34 Seminar: Data Mining on Biometric identification

How To Get A Computer Engineering Degree

Chapter 5 Business Intelligence: Data Warehousing, Data Acquisition, Data Mining, Business Analytics, and Visualization

Least Squares Estimation

Data Mining and Machine Learning in Bioinformatics

UF EDGE brings the classroom to you with online, worldwide course delivery!

Masters in Artificial Intelligence

Graduate Co-op Students Information Manual. Department of Computer Science. Faculty of Science. University of Regina

Overview. Swarms in nature. Fish, birds, ants, termites, Introduction to swarm intelligence principles Particle Swarm Optimization (PSO)

Supervised and unsupervised learning - 1

School of Computer Science

Data Mining for Customer Service Support. Senioritis Seminar Presentation Megan Boice Jay Carter Nick Linke KC Tobin

ISSUES IN RULE BASED KNOWLEDGE DISCOVERING PROCESS

BEHAVIOR BASED CREDIT CARD FRAUD DETECTION USING SUPPORT VECTOR MACHINES

Machine Learning and Data Mining. Fundamentals, robotics, recognition

Interdisciplinary Master s study program in Computer Science and Mathematics

Government of Russian Federation. Faculty of Computer Science School of Data Analysis and Artificial Intelligence

School of Computer Science

EXECUTIVE SUPPORT SYSTEMS (ESS) STRATEGIC INFORMATION SYSTEM DESIGNED FOR UNSTRUCTURED DECISION MAKING THROUGH ADVANCED GRAPHICS AND COMMUNICATIONS *

A Partially Supervised Metric Multidimensional Scaling Algorithm for Textual Data Visualization

Analecta Vol. 8, No. 2 ISSN

Machine Learning/Data Mining for Cancer Genomics

Defending Networks with Incomplete Information: A Machine Learning Approach. Alexandre

Pattern Analysis. Logistic Regression. 12. Mai Joachim Hornegger. Chair of Pattern Recognition Erlangen University

MASTER OF SCIENCE IN Computing & Data Analytics. (M.Sc. CDA)

Faculty of Science School of Mathematics and Statistics

A.I. in health informatics lecture 1 introduction & stuff kevin small & byron wallace

Software Engineering and Service Design: courses in ITMO University

Masters in Human Computer Interaction

The Challenge of Handling Large Data Sets within your Measurement System

Business Analytics. 0. Overview. Lars Schmidt-Thieme. Information Systems and Machine Learning Lab (ISMLL) University of Hildesheim, Germany

Masters in Advanced Computer Science

A Review of Data Mining Techniques

Doctor of Philosophy in Computer Science

MS1b Statistical Data Mining

Using Artificial Intelligence to Manage Big Data for Litigation

CPSC 340: Machine Learning and Data Mining. Mark Schmidt University of British Columbia Fall 2015

Introduction to Data Mining and Machine Learning Techniques. Iza Moise, Evangelos Pournaras, Dirk Helbing

Fuzzy Cognitive Map for Software Testing Using Artificial Intelligence Techniques

Visualization of Breast Cancer Data by SOM Component Planes

Programming Languages

Computational Science and Informatics (Data Science) Programs at GMU

CHARACTERISTICS IN FLIGHT DATA ESTIMATION WITH LOGISTIC REGRESSION AND SUPPORT VECTOR MACHINES

Unsupervised and supervised dimension reduction: Algorithms and connections

Data Mining course Master in Information Technologies Enginyeria Informàtica Tomàs Aluja. LIAM EIO. UPC Lluis Belanche LSI. UPC

Generalizing Random Forests Principles to other Methods: Random MultiNomial Logit, Random Naive Bayes, Anita Prinzie & Dirk Van den Poel

Linear Algebra Methods for Data Mining

Computer Information Systems

Classification algorithm in Data mining: An Overview

Draft dpt for MEng Electronics and Computer Science

Statistics for BIG data

Chapter 5. Warehousing, Data Acquisition, Data. Visualization

Data visualization and dimensionality reduction using kernel maps with a reference point

Data Mining, Predictive Analytics with Microsoft Analysis Services and Excel PowerPivot

Decompose Error Rate into components, some of which can be measured on unlabeled data

Course 395: Machine Learning

What is Artificial Intelligence?

Sharing the experiences of teaching business analytics in a University course

Artificial Intelligence and Politecnico di Milano. Presented by Matteo Matteucci

Foundations of Artificial Intelligence. Introduction to Data Mining

Semi-Supervised and Unsupervised Machine Learning. Novel Strategies

Learning outcomes. Knowledge and understanding. Competence and skills

Transcription:

9.54 class 2 Types of learning + biophysical mechanisms of plasticity Shimon Ullman + Tomaso Poggio Danny Harari + Daniel Zysman + Darren Seibert 9.54, fall semester 24

An introduction to basic Machine Learning! oncepts Algorithms 9.54, fall semester 24

oncepts and algorithms in Machine Learning References: T. Hastie, R. Tibshirani, and J. Friedman. The Elements of Statistical Learning: Prediction, Inference and Data Mining. Second Edition, Springer Verlag, 29 (available for free from the author's website). Further readings : T. Poggio and S. Smale. The Mathematics of Learning: Dealing with Data. Notices of the AMS, 23 Pedro Domingos. A few useful things to know about machine learning. ommunications of the AM AM Homepage archive. Volume 55 Issue, October 22 Pages 78-87..! Useful Links MIT 9.52: Statistical Learning Theory and Applications, Fall 23 (http://www.mit.edu/~9.52/). Stanford S229 Machine Learning Autumn 23 (http://cs229.stanford.edu). See also the oursera version (https://www.coursera.org/course/ml).! 9.54, fall semester 24

4

5

~2 years ago in my group 6

~7 years ago in my group 7

8

9

Levels of Understanding (977--22)! Evolution! Learning and development! omputation! Algorithms! Wetware, hardware, circuits and components

uild intelligent machines Understand Learning Learning is the gateway to understanding intelligence

(Artifical) Intelligence: A Working Definition Turing test Ingredients for AI! natural language processing! knowledge representation! automated reasoning! machine learning! computer vision! robotics to manipulate

Exciting Days for AI Artificial intelligence systems have recently seen some striking successes Learning

ML and AI Machine Learning systems are trained on examples rather than being programmed

Data Unsupervised Feature Learning lassic Supervised Learning Learning Structure.. Dynamic Learning Machine Learning Problems and Approaches!

asic Setting: lassification (x,y ),...,(x n,y n ) x i 2 R p and y i 2 Y = {, }, i =,...,n X n = x......... x p..... x n......... x p n A Y n = y. y n A

Genomics n patients p gene expression measurements... ; X n = x......... x p..... x n......... x p n A; Y n = y. y n A...

Text lassification: ag of Words X n = x......... x p..... x n......... x p n A

Image lassification......

Image lassification X n = x......... x p..... x n......... x p n A

Video lassification: Action Recognition

From classification to regression (x,y ),...,(x n,y n ) x i 2 R D and y i 2 Y = {, }, i =,...,n y i 2 Y 2 R, i =,...,n housing prices Living area (feet 2 ) Price ($s) 24 4 6 33 24 369 46 232 3 54.. price (in $) 9 8 7 6 5 4 3 2 Living area (feet 2 ) #bedrooms Price ($s) 24 3 4 6 3 33 24 3 369 46 2 232 3 4 54... 5 5 2 25 3 35 4 45 5 square feet y i = f(x i )+ " i, > e.g. f(x) =w T x, " i N(, )

! Learning Problems Supervised Learning Unsupervised Reinforcement Learning...! Learning types atch Learning Online Active...! Machine Learning: Problems and Approaches

Variations on a Theme (x,y ),...,(x n,y n ) Multiclass: x i 2 R D and y i 2 Y = {,...,T}, i =,...,n Multitask: x i 2 R D and y i 2 R T, i =,...,n Learning a similarity function (x,x ; y, ), (x,x 2 ; y,2 )...,(x n,x n ; y n,n ) x j,x i 2 R D and y i,j 2 [, ], j, i =,...,n

Semisupervised Learning n = x......... x p..... x n......... x p n A Y n = ; y. y n A [ X u = x......... x p..... x u......... x p u A

Semisupervised Learning n = x......... x p..... x n......... x p n A Y n = ; y. y n A [ X u = x......... x p..... x u......... x p u A

Semisupervised Learning n = x......... x p..... x n......... x p n A Y n = ; y. y n A [ X u = x......... x p..... x u......... x p u A

Semisupervised Learning n = x......... x p..... x n......... x p n A Y n = ; y. y n A [ X u = x......... x p..... x u......... x p u A

Semisupervised Learning n = x......... x p..... x n......... x p n A Y n = ; y. y n A [ X u = x......... x p..... x u......... x p u A

Semisupervised Learning

Semisupervised Learning n = x......... x p..... x n......... x p n A Y n = ; y. y n A [ X u = x......... x p..... x u......... x p u A Manifold Learning

Unsupervised Learning X n = x......... x p..... x n......... x p n A Y n = y. y n A Given x,...,x n Goal: Extract patterns... lustering, k-means! Vector Quantization! Dimensionality reduction!...

! Learning Problems Supervised Learning Unsupervised Reinforcement Learning...!! Learning types atch Learning Online Active...! Machine Learning: Problems and Approaches

! Learning Problems Supervised Learning Unsupervised Reinforcement Learning...! Learning types atch Learning Online Active...! Machine Learning: Problems and Approaches

Online/Incremental Learning (x,y ),...,(x n,y n ) f f (x,y ) (x 2,y 2 ) f 2... (x n,y n ) f n (x,y ) f...

The final estimator is constructed assembling the estimate away from th boundary obtained in the preview step with the estimate in the vicinity of th f Active Learning Learner can query points (x,y ) f (x 2,y 2 ) f 2... (a) (b) (x n,y n ) f n... (c) (d) Figure 8.7. The two step procedure for d =2:(a)InitialunprunedRDPandn/2 sample (b) Preview step RDP. Note that the cell with the arrow was pruned, but it contains a part of th boundary. (c) Additional sampling for the refinement step. (d) Refinement step.

Summary ``Learning is the acquisition of knowledge or skills through study, experience, or being taught We look for systems that are trained, rather than programmed, to perform a task! Learning from examples is a unifying paradigm in learning:!