My Wild & Crazy Idea: Causal Computational Learning Theory Scott Aaronson But first: what ever happened to my WACI from a few years ago: a Web 2.0 mathematics.

Slides:



Advertisements
Similar presentations
Types of Science:.
Advertisements

How Much Information Is In Entangled Quantum States? Scott Aaronson MIT |
New Evidence That Quantum Mechanics Is Hard to Simulate on Classical Computers Scott Aaronson Parts based on joint work with Alex Arkhipov.
Pretty-Good Tomography Scott Aaronson MIT. Theres a problem… To do tomography on an entangled state of n qubits, we need exp(n) measurements Does this.
The Complexity of Agreement A 100% Quantum-Free Talk Scott Aaronson MIT.
Steve Weber UC Berkeley What are the goals of theory? To understand To predict To influence To control.
The influence of domain priors on intervention strategy Neil Bramley.
BAYESIAN NETWORKS Ivan Bratko Faculty of Computer and Information Sc. University of Ljubljana.
Great Theoretical Ideas in Computer Science for Some.
CSE 5522: Survey of Artificial Intelligence II: Advanced Techniques Instructor: Alan Ritter TA: Fan Yang.
1 Polynomial Time Probabilistic Learning of a Subclass of Linear Languages with Queries Yasuhiro TAJIMA, Yoshiyuki KOTANI Tokyo Univ. of Agri. & Tech.
Probabilities Random Number Generators –Actually pseudo-random –Seed Same sequence from same seed Often time is used. Many examples on web. Custom random.
Probably Approximately Correct Learning Yongsub Lim Applied Algorithm Laboratory KAIST.
1 Learning Entity Specific Models Stefan Niculescu Carnegie Mellon University November, 2003.
Probably Approximately Correct Model (PAC)
1 Trends in Mathematics: How could they Change Education? László Lovász Eötvös Loránd University Budapest.
Machine Learning Theory Maria-Florina Balcan Lecture 1, Jan. 12 th 2010.
Quiz 4: Mean: 7.0/8.0 (= 88%) Median: 7.5/8.0 (= 94%)
Machine Learning Theory Maria-Florina (Nina) Balcan Lecture 1, August 23 rd 2011.
1 CS 178H Introduction to Computer Science Research What is CS Research?
Tennessee Technological University1 The Scientific Importance of Big Data Xia Li Tennessee Technological University.
COMP3503 Intro to Inductive Modeling
Midterm Review Rao Vemuri 16 Oct Posing a Machine Learning Problem Experience Table – Each row is an instance – Each column is an attribute/feature.
Machine Learning1 Machine Learning: Summary Greg Grudic CSCI-4830.
Learning Structure in Bayes Nets (Typically also learn CPTs here) Given the set of random variables (features), the space of all possible networks.
Introduction: Why statistics? Petter Mostad
Bayesian Learning By Porchelvi Vijayakumar. Cognitive Science Current Problem: How do children learn and how do they get it right?
Chapter 1.1 Review. 1. What are the seven basic areas of physics study?
Introduction to Data Mining Group Members: Karim C. El-Khazen Pascal Suria Lin Gui Philsou Lee Xiaoting Niu.
INTRODUCTION TO DATA MINING MIS2502 Data Analytics.
DAY 2: THE SCIENTIFIC METHOD Lakki Chandrasekaran August 19,
I'm thinking of a number. 12 is a factor of my number. What other factors MUST my number have?
Bayesian Classification. Bayesian Classification: Why? A statistical classifier: performs probabilistic prediction, i.e., predicts class membership probabilities.
Physics Fluctuomatics (Tohoku University) 1 Physical Fluctuomatics 12th Bayesian network and belief propagation in statistical inference Kazuyuki Tanaka.
Problem Limited number of experimental replications. Postgenomic data intrinsically noisy. Poor network reconstruction.
CHAPTER 5 Probability Theory (continued) Introduction to Bayesian Networks.
 Science has a standard way to test an idea  Cause and effect  What does that means?  That everything that happens in this world is because of the.
"Classical" Inference. Two simple inference scenarios Question 1: Are we in world A or world B?
THE NATURE OF SCIENCE WHAT IS SCIENCE?. REAL VS FAKE SCIENCE Science is an organized way of gathering and analyzing evidence about the natural world Pseudoscience.
Tools of Environmental Scientist Chapter 2.  Scire (latin)  to know What is Science?
MIS2502: Data Analytics Advanced Analytics - Introduction.
Goal of Learning Algorithms  The early learning algorithms were designed to find such an accurate fit to the data.  A classifier is said to be consistent.
Scientific Method Chapter 1-1. What is Science?  Science – organized way of gathering and analyzing evidence about the natural world  Described as a.
Environmental Science October 27 and 28. Welcome! Paradigm (noun) A theory or a group of ideas about how something should be done, made, or thought about.
Monday, January 11,  INSTRUCTORS  STUDENTS:  Name?  Class?  Hometown?  Major?  Background: Math? Computers? Statistics?  Why did you take.
CS 8751 ML & KDDComputational Learning Theory1 Notions of interest: efficiency, accuracy, complexity Probably, Approximately Correct (PAC) Learning Agnostic.
Pseudo-random generators Talk for Amnon ’ s seminar.
Richard W. Hamming Learning to Learn The Art of Doing Science and Engineering Session 24: Quantum Mechanics Learning to Learn The Art of Doing Science.
1 Guess the Covered Word Goal 1 EOC Review 2 Scientific Method A process that guides the search for answers to a question.
Machine Learning Chapter 7. Computational Learning Theory Tom M. Mitchell.
Neural Codes. Neuronal codes Spiking models: Hodgkin Huxley Model (brief repetition) Reduction of the HH-Model to two dimensions (general) FitzHugh-Nagumo.
1 Neural Codes. 2 Neuronal Codes – Action potentials as the elementary units voltage clamp from a brain cell of a fly.
Integrative Genomics I BME 230. Probabilistic Networks Incorporate uncertainty explicitly Capture sparseness of wiring Incorporate multiple kinds of data.
Table of Contents 2.4 How Does Scientific Knowledge Developed? Why Do Scientist use Models? What is a system? How are Models of Systems Used? Models as.
CS623: Introduction to Computing with Neural Nets (lecture-18) Pushpak Bhattacharyya Computer Science and Engineering Department IIT Bombay.
Graduate School of Information Sciences, Tohoku University
Probability and Statistics
MIS2502: Data Analytics Advanced Analytics - Introduction
Chapter 7: Sampling Distributions
Experimental Inquiry Template.
Large Scale Data Integration
CHAPTER 7 BAYESIAN NETWORK INDEPENDENCE BAYESIAN NETWORK INFERENCE MACHINE LEARNING ISSUES.
Computational Learning Theory
Computational Learning Theory
CS 188: Artificial Intelligence Fall 2007
CS 188: Artificial Intelligence Fall 2008
Quantum Computing and the Quest for Quantum Computational Supremacy
Graduate School of Information Sciences, Tohoku University
Graduate School of Information Sciences, Tohoku University
Graduate School of Information Sciences, Tohoku University
Presentation transcript:

My Wild & Crazy Idea: Causal Computational Learning Theory Scott Aaronson But first: what ever happened to my WACI from a few years ago: a Web 2.0 mathematics discussion site and conjecture/theorem repository? There now exists such a site, Mathoverflow.net, which is everything I hoped for and more. In just ~2 years, its noticeably changed the practice of mathematics. I had nothing to do with its creation.

PAC (Probabilistically Approximately Correct) Learning Computational complexity theory meets statistics PAC-learning is a hugely successful modelbut like most statistics, it doesnt care about the distinction between correlation and cause Given a collection of labeled examples (x 1,f(x 1 )),…,(x m,f(x m )) drawn independently from some unknown distribution D, problem is to output a hypothesis h such that h(x)=f(x) for most x~D with high probability

The result: PAC explains how banks predict who will repay their loans, but not how Einstein predicted the bending of starlight by the sun vs To predict what will happen in novel situations, you need to know something about causal mechanismswhich often requires controlled experiments (together with prior knowledge about temporal direction, autonomy of subsystems, etc.) Best theory of causality we currently have: Judea Pearls do-calculus

My WACI Challenge for Theory Traditional statistics : PAC-learning :: Pearls do-calculus : what? Potential applications: Debugging, reconstructing gene regulatory networks… Existing work in the direction Im talking about: - PAC-learning with membership and equivalence queries - Angluin, Aspnes, Chen, Wu: Learning a circuit by injecting values, STOC Pearls IC algorithm - Leakage-resilience in cryptography