Commonsense Reasoning and Argumentation 14/15 HC 10: Structured argumentation (3) Henry Prakken 16 March 2015.

Slides:



Advertisements
Similar presentations
Some important properties Lectures of Prof. Doron Peled, Bar Ilan University.
Advertisements

Discrete Math Methods of proof 1.
Introduction to Proofs
Argumentation Based on the material due to P. M. Dung, R.A. Kowalski et al.
Logic Programming Automated Reasoning in practice.
Commonsense Reasoning and Argumentation 14/15 HC 8 Structured argumentation (1) Henry Prakken March 2, 2015.
1 Default Reasoning Q: How many wheels does John’s car have? A: Four (by default) The conclusion is withdrawn if one is supplied with the information that.
Computational Models for Argumentation in MAS
Commonsense Reasoning and Argumentation 14/15 HC 9 Structured argumentation (2) Henry Prakken March 4, 2015.
On the structure of arguments, and what it means for dialogue Henry Prakken COMMA-08 Toulouse,
The Logic of Intelligence Pei Wang Department of Computer and Information Sciences Temple University.
Default Reasoning the problem: in FOL, universally-quantified rules cannot have exceptions –  x bird(x)  can_fly(x) –bird(tweety) –bird(opus)  can_fly(opus)
Default Reasoning By Naval Chopra( ) ‏ Pranay Bhatia ( ) ‏ Pradeep Kumar(07D05020) ‏ Siddharth Chinoy(07D05005) ‏ Vaibhav Chhimpa(07D05011)
Deductive Arguments: Categorical Logic
Legal Argumentation 2 Henry Prakken March 28, 2013.
1 DCP 1172 Introduction to Artificial Intelligence Chang-Sheng Chen Topics Covered: Introduction to Nonmonotonic Logic.
Argumentation Logics Lecture 5: Argumentation with structured arguments (1) argument structure Henry Prakken Chongqing June 2, 2010.
Proofs, Recursion and Analysis of Algorithms Mathematical Structures for Computer Science Chapter 2 Copyright © 2006 W.H. Freeman & Co.MSCS SlidesProofs,
Argumentation Logics Lecture 1: Introduction Henry Prakken Chongqing May 26, 2010.
BIRDS FLY. is a bird Birds fly Tweety is a bird Tweety flies DEFEASIBLE NON-MONOTONIC PRESUMPTIVE?
Argumentation Logics Lecture 7: Argumentation with structured arguments (3) Rationality postulates, Self-defeat Henry Prakken Chongqing June 4, 2010.
Commonsense Reasoning and Argumentation 14/15 HC 12 Dynamics of Argumentation (1) Henry Prakken March 23, 2015.
Relative Expressiveness of Defeasible Logics II Michael Maher.
Some problems with modelling preferences in abstract argumentation Henry Prakken Luxemburg 2 April 2012.
The Argument Mapping Tool of the Carneades Argumentation System DIAGRAMMING EVIDENCE: VISUALIZING CONNECTIONS IN SCIENCE AND HUMANITIES’ DIAGRAMMING EVIDENCE:
Commonsense Reasoning and Argumentation 14/15 HC 13: Dialogue Systems for Argumentation (1) Henry Prakken 25 March 2015.
Argumentation in Artificial Intelligence Henry Prakken Lissabon, Portugal December 11, 2009.
FINDING THE LOGIC OF ARGUMENTATION Douglas Walton CRRAR Coimbra, March 24, 2011.
1 OSCAR: An Architecture for Generally Intelligent Agents John L. Pollock Philosophy and Cognitive Science University of Arizona
Outline Recap Knowledge Representation I Textbook: Chapters 6, 7, 9 and 10.
Reasoning with testimony Argumentation vs. Explanatory Coherence Floris Bex - University of Groningen Henry Prakken - University of Groningen - Utrecht.
Auto-Epistemic Logic Proposed by Moore (1985) Contemplates reflection on self knowledge (auto-epistemic) Allows for representing knowledge not just about.
Argumentation Logics Lecture 6: Argumentation with structured arguments (2) Attack, defeat, preferences Henry Prakken Chongqing June 3, 2010.
Argumentation Henry Prakken SIKS Basic Course Learning and Reasoning May 26 th, 2009.
So far we have learned about:
ARTIFICIAL INTELLIGENCE TECHNIQUES Knowledge Processing 1.
Argumentation Logics Lecture 7: Argumentation with structured arguments (3) Henry Prakken Chongqing June 4, 2010.
Argumentation Logics Lecture 6: Argumentation with structured arguments (2) Attack, defeat, preferences Henry Prakken Chongqing June 3, 2010.
Argumentation Logics Lecture 3: Abstract argumentation semantics (3) Henry Prakken Chongqing May 28, 2010.
Argumentation Logics Lecture 4: Games for abstract argumentation Henry Prakken Chongqing June 1, 2010.
Argumentation Logics Lecture 1: Introduction Henry Prakken Chongqing May 26, 2010.
Proof by Deduction. Deductions and Formal Proofs A deduction is a sequence of logic statements, each of which is known or assumed to be true A formal.
EE1J2 – Discrete Maths Lecture 5 Analysis of arguments (continued) More example proofs Formalisation of arguments in natural language Proof by contradiction.
Argumentation Logics Lecture 5: Argumentation with structured arguments (1) argument structure Henry Prakken Chongqing June 2, 2010.
Intro to AI Fall 2002 © L. Joskowicz 1 Introduction to Artificial Intelligence LECTURE 11: Nonmonotonic Reasoning Motivation: beyond FOL + resolution Closed-world.
Belief Revision Lecture 1: AGM April 1, 2004 Gregory Wheeler
Henry Prakken August 23, 2013 NorMas 2013 Argumentation about Norms.
I NTRO TO L OGIC Dr Shlomo Hershkop March
Introduction to formal models of argumentation
Legal Argumentation 3 Henry Prakken April 4, 2013.
Commonsense Reasoning and Argumentation 14/15 HC 14: Dialogue systems for argumentation (2) Henry Prakken 30 March 2015.
LECTURE 17 THE MODAL ONTOLOGICAL ARGUMENT (A VARIANT OF HARTSHORNE’S VERSION)
Knowledge Processing 1. Aims of session  Introduce types of reasoning  Deterministic  Propositional logic  Predicate logic.
On the Semantics of Argumentation 1 Antonis Kakas Francesca Toni Paolo Mancarella Department of Computer Science Department of Computing University of.
CS104:Discrete Structures Chapter 2: Proof Techniques.
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.
Belief dynamics and defeasible argumentation in rational agents M. A. Falappa - A. J. García G. R. Simari Artificial Intelligence Research and Development.
1 Propositional Proofs 1. Problem 2 Deduction In deduction, the conclusion is true whenever the premises are true.  Premise: p Conclusion: (p ∨ q) 
Chapter 1 Logic and proofs
October 19th, 2007L. M. Pereira and A. M. Pinto1 Approved Models for Normal Logic Programs Luís Moniz Pereira and Alexandre Miguel Pinto Centre for Artificial.
2. The Logic of Compound Statements Summary
Henry Prakken & Giovanni Sartor July 16, 2012
Contradiction-tolerant TMS
Henry Prakken Guangzhou (China) 10 April 2018
Henry Prakken COMMA 2016 Berlin-Potsdam September 15th, 2016
Henry Prakken February 23, 2018
Propositional Logic.
Logical Entailment Computational Logic Lecture 3
Based on the material due to P. M. Dung, R.A. Kowalski et al.
Logic Logic is a discipline that studies the principles and methods used to construct valid arguments. An argument is a related sequence of statements.
Presentation transcript:

Commonsense Reasoning and Argumentation 14/15 HC 10: Structured argumentation (3) Henry Prakken 16 March 2015

Overview More about rationality postulates Related research The need for defeasible rules

3 Subtleties concerning rebuttals (1) d1: Ring  Married d2: Party animal  Bachelor s1: Bachelor  ¬Married K n : Ring, Party animal d2 < d1

4 Subtleties concerning rebuttals (2) d1: Ring  Married d2: Party animal  Bachelor s1: Bachelor  ¬Married s2: Married  ¬Bachelor K n : Ring, Party animal d2 < d1

5 Subtleties concerning rebuttals (3) d1: Ring  Married d2: Party animal  Bachelor s1: Bachelor  ¬Married s2: Married  ¬Bachelor K n : Ring, Party animal

6 Subtleties concerning rebuttals (4) R d = { ,      } R s = all deductively valid inference rules K n : d1: Ring  Married d2: Party animal  Bachelor n1: Bachelor  ¬Married Ring, Party animal

7 Argumentation systems (with generalised negation) An argumentation system is a tuple AS = ( L, -, R,n) where: L is a logical language - is a contrariness function from L to 2 L R = R s  R d is a set of strict and defeasible inference rules n: R d  L is a naming convention for defeasible rules

8 Generalised negation The – function generalises negation. If   - (  ) then: if   - (  ) then  is a contrary of  ; if   - (  ) then  and  are contradictories We write -  = ¬  if  does not start with a negation -  =  if is of the form ¬ 

9 Attack and defeat (the general case) A undermines B (on  ) if Conc(A) = -  for some   Prem(B )/ K n ; A rebuts B (on B’ ) if Conc(A) = -Conc(B’ ) for some B’  Sub(B ) with a defeasible top rule; A undercuts B (on B’ ) if Conc(A) = -n(r) ’for some B’  Sub(B ) with defeasible top rule r A contrary-undermines/rebuts B (on  /B’ ) if Conc(A) is a contrary of  / Conc(B ’) A defeats B iff for some B’ A undermines B on  and either A contrary-undermines B’ on  or not A < a  ; or A rebuts B on B’ and either A contrary-rebuts B’ or not A < a B’ ; or A undercuts B on B’

10 Consistency in ASPIC+ (with generalised negation) For any S  L S is directly consistent iff S does not contain two formulas  and –(  ) The strict closure Cl(S) of S is S + everything derivable from S with only R s. S is indirectly consistent iff Cl(S) is directly consistent. Parametrised by choice of strict rules

11 Rationality postulates for ASPIC+ (with generalised negation) Closure under subarguments always satisfied Direct and indirect consistency: without preferences satisfied if R s closed under transposition or AS closed under contraposition; and K n is indirectly consistent; and AT is `well-formed’ with preferences satisfied if in addition  is ‘reasonable’ Weakest- and last link ordering are reasonable AT is well-formed if: If  is a contrary of  then (1)   K n and (2)  is not the consequent of a strict rule

Relation with other work (1) Assumption-based argumentation (Dung, Kowalski, Toni...) is special case of ASPIC+ (with generalised negation) with Only ordinary premises Only strict inference rules All arguments of equal priority …

Reduction of ASPIC+ defeasible rules to ABA rules (Dung & Thang, JAIR 2014) Assumptions: L consists of literals No preferences No rebuttals of undercutters p 1, …, p n  q becomes d i, p 1, …, p n,not¬q  q where: d i = n(p 1, …, p n  q) d i, not¬q are assumptions  = - (not  ),  = - (¬  ), ¬  = - (  ) 1-1 correspondence between grounded, preferred and stable extensions of ASPIC+ and ABA

14 From defeasible to strict rules: example r1: Quaker  Pacifist r2: Republican  ¬Pacifist Pacifist Quaker  Pacifist Republican r1 r2

15 From defeasible to strict rules: example s1: Appl(s1), Quaker, not¬Pacifist  Pacifist s2: Appl(s2), Republican, notPacifist  ¬Pacifist Pacifist QuakerAppl(s1)not¬Pacifist ¬Pacifist RepublicannotPacifistAppl(s2)

Can ASPIC+ preferences be reduced to ABA assumptions? d1: Bird  Flies d2: Penguin  ¬Flies d1 < d2 Becomes d1: Bird, not Penguin  Flies d2: Penguin  ¬Flies Only works in special cases, e.g. not with weakest-link ordering

Classical argumentation (Besnard & Hunter, …) Given L a propositional logical language and |- standard- logical consequence over L : An argument is a pair (S,p) such that S  L and p  L S |- p S is consistent No S’  S is such that S’ |- p Various notions of attack, e.g.: “Direct defeat”: argument (S,p) attacks argument (S’,p’) iff p |- ¬q for some q  S’ “Direct undercut”: argument (S,p) attacks argument (S’,p’) iff p |- ¬q and ¬q |- p for some q  S’ Only these two attacks satisfy consistency.

Relation with other work (2) Two variants of classical argumentation with premise attack (Amgoud & Cayrol, Besnard & Hunter) are special case of ASPIC+ with Only ordinary premises Only strict inference rules (all valid propositional or first-order inferences from finite sets) - = ¬ No preferences Arguments must have classically consistent premises …

Results on classical argumentation (Cayrol 1995; Amgoud & Besnard 2013) In classical argumentation with premise attack, only ordinary premises and no preferences: Preferred and stable extensions and maximal conflict-free sets coincide with maximal consistent subsets of the knowledge base So p is defensible iff there exists an argument for p The grounded extension coincides with the intersection of all maximal consistent subsets of the knowledge base So p is justified iff there exists an argument for p without counterargument Lindebaum’s lemma: Every consistent set is contained in a maximal consistent set

20 Modelling default reasoning in classical argumentation Quakers are usually pacifist Republicans are usually not pacifist Nixon was a quaker and a republican

21 A modelling in classical logic K p : Quaker  Pacifist Republican   ¬Pacifist Quaker, Republican Pacifist Quaker Quaker  Pacifist ¬Pacifist Republican Republican  ¬Pacifist ¬(Quaker  Pacifist)

22 A modelling in classical logic K n : Quaker & ¬Ab1  Pacifist Republican & ¬Ab2   ¬Pacifist Quaker, Republican K p : ¬Ab1, ¬Ab2 (attackable) Pacifist Quaker¬Ab1 ¬Pacifist ¬Ab2Republican Quaker & ¬Ab1  Pacifist Republican & ¬Ab2  ¬Pacifist Ab1

23 A modelling in classical logic Pacifist Quaker¬Ab1 ¬Pacifist ¬Ab2Republican Quaker & ¬Ab1  Pacifist Republican & ¬Ab2  ¬Pacifist Ab1 Ab2

Can defeasible reasoning be reduced to plausible reasoning? Is it natural to reduce all forms of attack to premise attack? My answer: no In classical argumentation: can the material implication represent defaults? My answer: no

Default contraposition in classical argumentation Heterosexuals are normally married. John is not married Assume when possible that things are normal What can we conclude about John’s sexual orientation?

Default contraposition in classical argumentation Heterosexuals are normally married H & ¬Ab  M John is not married (¬M) Assume when possible that things are normal ¬Ab The first default implies that non-married people are normally not heterosexual ¬M & ¬Ab  ¬H So John is not heterosexual

Default contraposition in classical argumentation (2) Men normally have no beard => Creatures with a beard are normally not men This type of sensor usually does not give false alarms => False alarms are usually not given by this type of sensor Witnesses interrogated by the police usually tell the truth => People interrogated by the police who do not speak the truth are usually not a witness Statisticians call these inferences “base rate fallacies”

The case of classical argumentation Birds usually fly Penguins usually don’t fly All penguins are birds Penguins are abnormal birds w.r.t. flying Tweety is a penguin

The case of classical argumentation Birds usually fly Bird & ¬Ab1  Flies Penguins usually don’t fly Penguin & ¬Ab2  ¬Flies All penguins are birds Penguin  Bird Penguins are abnormal birds w.r.t. flying Penguin  Ab1 Tweety is a penguin Penguin ¬Ab1 ¬Ab2

The case of classical argumentation Bird & ¬Ab1  Flies Penguin & ¬Ab2  ¬Flies Penguin  Bird Penguin  Ab1 Penguin ¬Ab1 ¬Ab2 Arguments: - for Flies using ¬Ab1 - for ¬Flies using ¬Ab2 KpKp KnKn

The case of classical argumentation Bird & ¬Ab1  Flies Penguin & ¬Ab2  ¬Flies Penguin  Bird Penguin  Ab1 Penguin ¬Ab1 ¬Ab2 Arguments: - for Flies using ¬Ab1 - for ¬Flies using ¬Ab2 - and for Ab1 and Ab2 But ¬Flies follows KpKp KnKn

The case of classical argumentation Bird & ¬Ab1  Flies Penguin & ¬Ab2  ¬Flies Penguin  Bird Penguin  Ab1 ObservedAsPenguin & ¬Ab3  Penguin ObservedAsPenguin ¬Ab1 ¬Ab2 ¬Ab3 Arguments: - for Flies using ¬Ab1 - for ¬Flies using ¬Ab2 and ¬Ab3 - for Penguin using ¬Ab3

The case of classical argumentation Bird & ¬Ab1  Flies Penguin & ¬Ab2  ¬Flies Penguin  Bird Penguin  Ab1 ObservedAsPenguin & ¬Ab3  Penguin ObservedAsPenguin ¬Ab1 ¬Ab2 ¬Ab3 Arguments: - for Flies using ¬Ab1 - for ¬Flies using ¬Ab2 and ¬Ab3 - for Penguin using ¬Ab3 - and for Ab1 and Ab2 and Ab3

The case of classical argumentation Bird & ¬Ab1  Flies Penguin & ¬Ab2  ¬Flies Penguin  Bird Penguin  Ab1 ObservedAsPenguin & ¬Ab3  Penguin ObservedAsPenguin ¬Ab1 ¬Ab2 ¬Ab3 Arguments: - for Flies using ¬Ab1 - for ¬Flies using ¬Ab2 and ¬Ab3 - for Penguin using ¬Ab3 - and for Ab1 and Ab2 and Ab3 ¬Ab3 > ¬Ab2 > ¬Ab1 makes ¬Flies follow But is this ordering natural?

35 Contraposition of legal rules r1: Snores  Misbehaves r2: Misbehaves  May be removed r3: Professor  ¬May be removed K: Snores, Professor r1 < r2, r1 < r3, r3 < r2 May be removed Misbehaves Snores  May be removed Professor r1 r2 r3 This is the intuitive outcome R3 < R2

36 Contraposition of legal rules r1: Snores  Misbehaves r2: Misbehaves  May be removed r3: Professor  ¬May be removed K: Snores, Professor r1 < r2, r1 < r3, r3 < r2 May be removed Misbehaves Snores  May be removed Professor r1 r2 r3 But with contraposition (and last or weakest link) we have this outcome

My conclusion Classical logic’s material implication is too strong for representing defeasible generalisations or legal rules => Models of legal argument (and many other kinds of argument) need defeasible inference rules Defeasible reasoning cannot be modelled as inconsistency handling in deductive logic John Pollock: Defeasible reasoning is the rule, deductive reasoning is the exception

Next lecture The lottery paradox Self-defeat and odd defeat loops Floating conclusions The need for dynamics