Lifted First-Order Probabilistic Inference Rodrigo de Salvo Braz, Eyal Amir and Dan Roth This research is supported by ARDA’s AQUAINT Program, by NSF grant.

Slides:



Advertisements
Similar presentations
1 MPE and Partial Inversion in Lifted Probabilistic Variable Elimination Rodrigo de Salvo Braz University of Illinois at Urbana-Champaign with Eyal Amir.
Advertisements

1 Classification using instance-based learning. 3 March, 2000Advanced Knowledge Management2 Introduction (lazy vs. eager learning) Notion of similarity.
Bayesian Network and Influence Diagram A Guide to Construction And Analysis.
1 CHAPTER 4 RELATIONAL ALGEBRA AND CALCULUS. 2 Introduction - We discuss here two mathematical formalisms which can be used as the basis for stating and.
Relational Schemas and Predicate Logic: Notation.
1 MPE and Partial Inversion in Lifted Probabilistic Variable Elimination Rodrigo de Salvo Braz University of Illinois at Urbana-Champaign with Eyal Amir.
BAYESIAN NETWORKS. Bayesian Network Motivation  We want a representation and reasoning system that is based on conditional independence  Compact yet.
The nature of Sign and sign/symbol distinction
Dynamic Bayesian Networks (DBNs)
UNCERTML - DESCRIBING AND COMMUNICATING UNCERTAINTY Matthew Williams
Knowledge Representation I Suppose I tell you the following... The Duck-Bill Platypus and the Echidna are the only two mammals that lay eggs. Only birds.
Review Markov Logic Networks Mathew Richardson Pedro Domingos Xinran(Sean) Luo, u
Causal and Bayesian Network (Chapter 2) Book: Bayesian Networks and Decision Graphs Author: Finn V. Jensen, Thomas D. Nielsen CSE 655 Probabilistic Reasoning.
Speaker:Benedict Fehringer Seminar:Probabilistic Models for Information Extraction by Dr. Martin Theobald and Maximilian Dylla Based on Richards, M., and.
Plan Recognition with Multi- Entity Bayesian Networks Kathryn Blackmond Laskey Department of Systems Engineering and Operations Research George Mason University.
A Probabilistic Framework for Information Integration and Retrieval on the Semantic Web by Livia Predoiu, Heiner Stuckenschmidt Institute of Computer Science,
10/24  Exam on 10/26 (Lei Tang and Will Cushing to proctor)
Cooperating Intelligent Systems Course review AIMA.
Learning with Bayesian Networks David Heckerman Presented by Colin Rickert.
1 Learning Entity Specific Models Stefan Niculescu Carnegie Mellon University November, 2003.
Big Ideas in Cmput366. Search Blind Search Iterative deepening Heuristic Search A* Local and Stochastic Search Randomized algorithm Constraint satisfaction.
Semantics For the Semantic Web: The Implicit, the Formal and The Powerful Amit Sheth, Cartic Ramakrishnan, Christopher Thomas CS751 Spring 2005 Presenter:
Constructing Belief Networks: Summary [[Decide on what sorts of queries you are interested in answering –This in turn dictates what factors to model in.
LEARNING FROM OBSERVATIONS Yılmaz KILIÇASLAN. Definition Learning takes place as the agent observes its interactions with the world and its own decision-making.
Let remember from the previous lesson what is Knowledge representation
10/22  Homework 3 returned; solutions posted  Homework 4 socket opened  Project 3 assigned  Mid-term on Wednesday  (Optional) Review session Tuesday.
Copyright © Cengage Learning. All rights reserved.
1 Bayesian Networks Chapter ; 14.4 CS 63 Adapted from slides by Tim Finin and Marie desJardins. Some material borrowed from Lise Getoor.
Knowledge representation methods جلسه سوم. KR is AI bottleneck The most important ingredient in any expert system is knowledge. The power of expert systems.
Data Modeling Using the Entity-Relationship Model
RDF (Resource Description Framework) Why?. XML XML is a metalanguage that allows users to define markup XML separates content and structure from formatting.
Relational Probability Models Brian Milch MIT 9.66 November 27, 2007.
1 MPE and Partial Inversion in Lifted Probabilistic Variable Elimination Rodrigo de Salvo Braz University of Illinois at Urbana-Champaign with Eyal Amir.
March 16 & 21, Csci 2111: Data and File Structures Week 9, Lectures 1 & 2 Indexed Sequential File Access and Prefix B+ Trees.
Bayesian Networks for Data Mining David Heckerman Microsoft Research (Data Mining and Knowledge Discovery 1, (1997))
Properties of Real Numbers. Sets In mathematics, a set is a collection of things Sets can be studies as a topic all on its own (known as set theory),
UNCERTML - DESCRIBING AND COMMUNICATING UNCERTAINTY WITHIN THE (SEMANTIC) WEB Matthew Williams
Lifted First-Order Probabilistic Inference Rodrigo de Salvo Braz SRI International joint work with Eyal Amir and Dan Roth.
1 Predicate (Relational) Logic 1. Introduction The propositional logic is not powerful enough to express certain types of relationship between propositions.
Formal Specification of Intrusion Signatures and Detection Rules By Jean-Philippe Pouzol and Mireille Ducassé 15 th IEEE Computer Security Foundations.
1 Lifted First-Order Probabilistic Inference Rodrigo de Salvo Braz University of Illinois at Urbana-Champaign with Eyal Amir and Dan Roth.
1 CMSC 671 Fall 2001 Class #21 – Tuesday, November 13.
Teaching to the “Big Ideas”: Moving beyond the standards Terry P. Vendlinski UCLA Graduate School of Education & Information Studies National Center for.
4. Particle Filtering For DBLOG PF, regular BLOG inference in each particle Open-Universe State Estimation with DBLOG Rodrigo de Salvo Braz*, Erik Sudderth,
Logical Agents Chapter 7. Outline Knowledge-based agents Logic in general Propositional (Boolean) logic Equivalence, validity, satisfiability.
The famous “sprinkler” example (J. Pearl, Probabilistic Reasoning in Intelligent Systems, 1988)
1 Use graphs and not pure logic Variables represented by nodes and dependencies by edges. Common in our language: “threads of thoughts”, “lines of reasoning”,
First-Order Probabilistic Inference Rodrigo de Salvo Braz.
Knowledge Discovery in Protected Vertical Information Dr. William Perrizo University Distinguished Professor of Computer Science North Dakota State University,
Review: What is a logic? A formal language –Syntax – what expressions are legal –Semantics – what legal expressions mean –Proof system – a way of manipulating.
Of 29 lecture 15: description logic - introduction.
An argument-based framework to model an agent's beliefs in a dynamic environment Marcela Capobianco Carlos I. Chesñevar Guillermo R. Simari Dept. of Computer.
First-Order Probabilistic Inference Rodrigo de Salvo Braz SRI International.
CIS750 – Seminar in Advanced Topics in Computer Science Advanced topics in databases – Multimedia Databases V. Megalooikonomou Link mining ( based on slides.
Probabilistic Reasoning Inference and Relational Bayesian Networks.
Lecture 14. Recap Problem Solving GA Simple GA Examples of Mutation and Crossover Application Areas.
Logical Agents. Outline Knowledge-based agents Logic in general - models and entailment Propositional (Boolean) logic Equivalence, validity, satisfiability.
1 Representing and Reasoning on XML Documents: A Description Logic Approach D. Calvanese, G. D. Giacomo, M. Lenzerini Presented by Daisy Yutao Guo University.
ece 720 intelligent web: ontology and beyond
Knowledge Representation and Reasoning
Learning Bayesian Network Models from Data
Knowledge Representation
Probabilistic Horn abduction and Bayesian Networks
Lifted First-Order Probabilistic Inference [de Salvo Braz, Amir, and Roth, 2005] Daniel Lowd 5/11/2005.
Probabilistic Reasoning; Network-based reasoning
Class #16 – Tuesday, October 26
Knowledge Representation
First Order Probabilistic Inference, by David Poole (2003)
CS 188: Artificial Intelligence Fall 2008
Habib Ullah qamar Mscs(se)
Presentation transcript:

Lifted First-Order Probabilistic Inference Rodrigo de Salvo Braz, Eyal Amir and Dan Roth This research is supported by ARDA’s AQUAINT Program, by NSF grant ITR-IIS and DAF grant FA (DARPA REAL program). AAAI’06, IJCAI’05 Goal A general system to reason with symbolic data in a probabilistic fashion Being able to describe objects of arbitrary structure: lists, trees, objects with various fields and refering to other objects; Being able to describe probabilistic facts. Previous approaches - Logic Powerful data representation, powerful knowledge declaration, but lack of probabilities; Prevents us to say things as simple as: “most birds fly”, “most capitalized words are proper nouns” or “rectangular objects are common”. Lifted First-Order Probabilistic Inference Allows for declaration of probabilistic knowledge (KB) that applies to many objects: Prob(properNoun(Word) | noun(Word), capitalized(Word), similarTo(Word, Word2), knownName(Word2)) = 0.7 holds for all words involved. Does not build BN in advance of inference. Instead, performs inference using KB items as necessary. Does not consider individual objects separately unless this is really necessary: similar to theorem-proving, but with probabilities. Example: given capitalized(Beautiful) and adjective(Beautiful), the algorithm does not have to consider the whole list of known names in order to decide that Beautiful is not a proper noun, because it is not a noun in the first place. An algorithm constructing a model before doing inference would build a node per known name! Example A group of people, each with data such as name, age, profession, as well as relationships between them: Prob(friends(X,Z) | friends(X,Y), friends(Y,Z)) = 0.6 Prob(age(Person) > 20 | degree(Person)) = 0.9. Key ideas Each atom in a rule is a parameterized random variable. noun(Word) actually stands for all its instance random variables noun(w 1 ), noun(w 2 ),..., word(w n ). We use a First-Order Variable Elimination algorithm, where often all instances of a parameterized random variable can be eliminated at once, regardless of the domain size, since they all have the same structure. This is an idea introduced by David Poole in 2003 that we formalized and call Inversion Elimination (IE). However, IE only works when these instances are independent of each other. For more general cases, we introduced Counting Elimination, which however does depend on the domain size. Contributions A First-Order Probabilistic Inference algorithm which: provides a language for both structured data and probabilistic information; takes advantage of first-order level information by eliminating many objects in a single step, as opposed to previous work. Comparing Lifted and Propositional Inferences We compare lifted methods (inversion and counting elimination) to their propositional counterparts (which produce identical numerical results), for P(sick(Person) | epidemics), P(death | sick(Person)) in (I) and P(p(X),p(Y),r) in (II). Example Web pages, with their words, subjects and links, and how they relate to each other. Prob(subject(Page,coffe) | wordIn(Page,java)) = 0.2 Prob(subject(Page,programming) | wordIn(Page,java)) = 0.8 Prob(subject(Page,Subj) | link(Page,Page2), subject(Page2, Subj)) = 0.7. Example Smoking facts in a large population: Prob(friends(X,Y)) = 0.1 Prob(smoker(X)) = 0.3 Prob(smoker(Y) | friends(X,Y), smoker(X)) = 0.6 Prob(cancer(X) | smoker(X)) = 0.6, X  john Prob(cancer(john)) = 0.01 A possible query would be: in a population of 40,000, what is the probability of a male having cancer? Previous approaches - Probabilistic models Powerful probabilistic capabilities in models such as Bayesian networks and Markov models, but lacks structured data such as lists, trees, etc. Cannot automatically apply knowledge to more than one object at the same time (for example, probabilistic properties of nouns to all nouns in a piece of text). Previous approach - Knowledge-driven Construction Constructs a propositional probabilistic model based on logic-like probabilistic knowledge; Builds model before starting inference, often building unnecessary parts. Where we are We have: predicates (relational structure); no individuals split unnecessarily. But we do not have yet: function symbols, so no data structures such as lists and trees; does not take full advantage of inference during construction. This will require approximation.