Implications as rules

Similar documents
Remarks on Dialogical Meaning: A Case Study Shahid Rahman 1 (Université de Lille, UMR: 8163, STL)

The tenure game. The tenure game. Winning strategies for the tenure game. Winning condition for the tenure game

18 Completeness and Compactness of First-Order Tableaux

On game semantics of the affine and intuitionistic logics (Extended abstract)

: Principles of Automated Reasoning and Decision Making Midterm

CMPUT 396 Tic-Tac-Toe Game

1.5 How Often Do Head and Tail Occur Equally Often?

Permutation Groups. Definition and Notation

Dynamic Games: Backward Induction and Subgame Perfection

Chapter 6 Rules of deductive inference Answers to select Getting familiar with exercises. Getting familiar with basic rules of inference.

5.4 Imperfect, Real-Time Decisions

ARTIFICIAL INTELLIGENCE (CS 370D)

Extensive Games with Perfect Information A Mini Tutorial

Games in Extensive Form

A paradox for supertask decision makers

Contents. MA 327/ECO 327 Introduction to Game Theory Fall 2017 Notes. 1 Wednesday, August Friday, August Monday, August 28 6

Formal Verification. Lecture 5: Computation Tree Logic (CTL)

2 person perfect information

THE PRINCIPLE OF PRESSURE IN CHESS. Deniz Yuret. MIT Articial Intelligence Laboratory. 545 Technology Square, Rm:825. Cambridge, MA 02139, USA

Logic and Artificial Intelligence Lecture 18

From a Ball Game to Incompleteness

Is everything stochastic?

22c181: Formal Methods in Software Engineering. The University of Iowa Spring Propositional Logic

Bricken Technologies Corporation Presentations: Bricken Technologies Corporation Corporate: Bricken Technologies Corporation Marketing:

Two Perspectives on Logic

A Historical Example One of the most famous problems in graph theory is the bridges of Konigsberg. The Real Koningsberg

Opponent Models and Knowledge Symmetry in Game-Tree Search

Dice Activities for Algebraic Thinking

Soundness and Completeness for Sentence Logic Derivations

Principle of Inclusion-Exclusion Notes

Progress in Computer Science and Applied Logic

CSEP 573 Adversarial Search & Logic and Reasoning

Awareness in Games, Awareness in Logic

Hackenbush. 1 Warm-ups. A Basic Problem

Final Exam : Constructive Logic. December 17, 2012

Foundations of AI. 5. Board Games. Search Strategies for Games, Games with Chance, State of the Art. Wolfram Burgard and Luc De Raedt SA-1

2359 (i.e. 11:59:00 pm) on 4/16/18 via Blackboard

Lecture 1, CS 2050, Intro Discrete Math for Computer Science

Strict Finitism Refuted? Ofra Magidor ( Preprint of paper forthcoming Proceedings of the Aristotelian Society 2007)

The study of probability is concerned with the likelihood of events occurring. Many situations can be analyzed using a simplified model of probability

A Linear-Logic Semantics for Constraint Handling Rules With Disjunction

1111: Linear Algebra I

Key stage 2 mathematics tasks for the more able Number slide solutions and what to look for

ADVERSARIAL SEARCH. Chapter 5

Extensive Games with Perfect Information. Start by restricting attention to games without simultaneous moves and without nature (no randomness).

The Odds Calculators: Partial simulations vs. compact formulas By Catalin Barboianu

Teacher s Notes. Problem of the Month: Courtney s Collection

Game Theory and Randomized Algorithms

Monday, February 2, Is assigned today. Answers due by noon on Monday, February 9, 2015.

Conversion Masters in IT (MIT) AI as Representation and Search. (Representation and Search Strategies) Lecture 002. Sandro Spina

Sequential games. We may play the dating game as a sequential game. In this case, one player, say Connie, makes a choice before the other.

Contents. Foundations of Artificial Intelligence. Problems. Why Board Games?

CIS/CSE 774 Principles of Distributed Access Control Exam 1 October 3, Points Possible. Total 60

Playing Games. Henry Z. Lo. June 23, We consider writing AI to play games with the following properties:

of the hypothesis, but it would not lead to a proof. P 1

Algorithms for Data Structures: Search for Games. Phillip Smith 27/11/13

Lecture 6: Metagaming

Computation. Philosophical Issues. Instructor: Viola Schiaffonati. March, 26 th 2018

A review of Reasoning About Rational Agents by Michael Wooldridge, MIT Press Gordon Beavers and Henry Hexmoor

Cognitive Illusions. Is this line straight? Can these lines possibly be parallel?

General Game Playing (GGP) Winter term 2013/ Summary

Planning and Optimization

Yale University Department of Computer Science

The topic for the third and final major portion of the course is Probability. We will aim to make sense of statements such as the following:

Your Name and ID. (a) ( 3 points) Breadth First Search is complete even if zero step-costs are allowed.

Finite games: finite number of players, finite number of possible actions, finite number of moves. Canusegametreetodepicttheextensiveform.

Dice Games and Stochastic Dynamic Programming

Combinatorics and Intuitive Probability

Symbolic Classification of General Two-Player Games

1. The chance of getting a flush in a 5-card poker hand is about 2 in 1000.

From: AAAI Technical Report FS Compilation copyright 1994, AAAI ( All rights reserved.

EXPLORING TIC-TAC-TOE VARIANTS

COUNTING AND PROBABILITY

SOLUTIONS TO PROBLEM SET 5. Section 9.1

Lecture 5: Subgame Perfect Equilibrium. November 1, 2006

Five-In-Row with Local Evaluation and Beam Search

Wednesday, February 1, 2017

THE TAYLOR EXPANSIONS OF tan x AND sec x

16 Alternating Groups

1. Introduction to Game Theory

Game-playing: DeepBlue and AlphaGo

Teaching and Learning Computation of Fractions Through Story Problems

Game Design Verification using Reinforcement Learning

and 6.855J. Network Simplex Animations

Human-Computer Interaction based on Discourse Modeling

Section Summary. Finite Probability Probabilities of Complements and Unions of Events Probabilistic Reasoning

Foundations of AI. 6. Board Games. Search Strategies for Games, Games with Chance, State of the Art

Objective of the Lecture

SF2972: Game theory. Mark Voorneveld, February 2, 2015

Logic and the Sizes of Sets

Automatic Generation of Web Interfaces from Discourse Models

CSE 100: RED-BLACK TREES

Weeks 3-4: Intro to Game Theory

Dynamic Programming in Real Life: A Two-Person Dice Game

37 Game Theory. Bebe b1 b2 b3. a Abe a a A Two-Person Zero-Sum Game

5.4 Imperfect, Real-Time Decisions

Propositional Calculus II: More Rules of Inference, Application to Additional Motivating Problems

Patterns and random permutations II

A Grid-Based Game Tree Evaluation System

UMBC CMSC 671 Midterm Exam 22 October 2012

Transcription:

DIPLEAP Wien 27.11.2010 p. 1 Implications as rules Thomas Piecha Peter Schroeder-Heister Wilhelm-Schickard-Institut für Informatik Universität Tübingen

DIPLEAP Wien 27.11.2010 p. 2 Philosophical / foundational perspective If we want to explain the meaning of implication, we need an elementary pre-logical notion at the structural level. Similar to pairing for conjunction, and schematic reasoning for universal quantification. For that we propose the notion of a rule. Proposal: At the logical level, an implication A B expresses a rule A B.

DIPLEAP Wien 27.11.2010 p. 3 The deductive meaning of implication Implications-as-rules corresponds to the deductive meaning of implication: A deduction Ạ. B justifies the rule A B and conversely, the rule A B justifies a deduction Ạ. B.

DIPLEAP Wien 27.11.2010 p. 4 Rules as primitives No specific syntactic notion of deduction is presupposed. Rules are primitive entities are rooted in everyday practice. Rules are prior to deductions. Rule following lies at the basis of cognitive activities.

DIPLEAP Wien 27.11.2010 p. 5 Implications as rules in natural deduction: modus ponens Modus ponens can be viewed as rule application: Read A B A B as A B A B The implications-as-rules view is inherent in natural deduction. Implication introduction = establishing a rule, modus ponens = applying a rule.

This expresses the notion of implications-as-rules in the sequent calculus. DIPLEAP Wien 27.11.2010 p. 6 Implications in the sequent calculus This differs from the view of implication in the sequent calculus. Gentzen s implication left schema: Γ A,B C Γ,,A B C This schema, which is based on a different intuition, also underlies the dialogical interpretation. As an alternative schema, I propose: Γ A Γ,A B B

Implications-as-rules from the database perspective: resolution Suppose the implication A B is available in our database. Then the goal B can be reduced to the goal A. More generally: Given a database (or logic program) A 1 B. A n B then the goal B can be reduced to any of the goals A i. This reduction is called resolution. Reasoning with respect to a database of implications means reading them as rules. DIPLEAP Wien 27.11.2010 p. 7

DIPLEAP Wien 27.11.2010 p. 8 Summary The understanding of implications as rules is philosophically fundamental psychologically elementary supported by the natural deduction view of reasoning supported by the database perspective Assuming an implication means: Putting it into a (virtual) database of rules, from which it can be applied in forward reasoning: modus ponens applied in backward reasoning: resolution

Dialogues A dialogue for a (b a) positions 0. P a (b a) 1. O a [0, attack] 2. P b a [1, defense] 3. O 2 [2, attack] 4. P a [3, defense] } {{ moves } Argumentation forms X and Y, where X Y, are variables for P and O. implication : assertion: X A B attack: Y A defense: X B conjunction : assertion: X A 1 A 2 attack: Y i (Y chooses i = 1 or i = 2) defense: X A i

Dialogues Dialogue (1) A dialogue is a sequence of moves (i) made alternatingly by P and O (ii) according to the argumentation forms, (iii) and P makes the first move. Dialogue (2) (D) P may assert an atomic formula only if it has been asserted by O before. (E) O can only react on the immediately preceding P-move. (plus some other conditions) A dialogue beginning with P A is called dialogue for the formula A. Asymmetry between proponent P and opponent O due to (D) and (E).

Strategies P wins a dialogue for a formula A if (i) the dialogue is finite, (ii) begins with the move P A and (iii) ends with a move of P such that O cannot make another move. Strategy A dialogue tree contains all possible dialogues for A as paths. P O O P P P O... A strategy for a formula A is a subtree S of the dialogue tree for A such that (i) S does not branch at even positions (i.e. at P-moves), (ii) S has as many nodes at odd positions as there are possible moves for O, (iii) all branches of S are dialogues for A won by P. P O O P P P O...

Strategies Example, strategy for (a b) ((b c) (a c)) 0. P (a b) ((b c) (a c)) 1. O a b [0, attack] 2. P (b c) (a c) [1, defense] 3. O b c [2, attack] 4. P a c [3, defense] 5. O a [4, attack] 6. P a [1, attack] 7. O b [6, defense] 8. P b [3, attack] 9. O c [8, defense] 10. P c [5, defense] A strategy for A is a proof of A.

Implications as Rules: Argumentation Forms assertion: attack: defense: O A B no attack (no defense) assertion: O A 1 A 2 attack: P i (i = 1 or 2) defense: O A i assertion: P A B question: O? choice: P A B P C only if O C (A B) before attack: O A defense: P B assertion: P A 1 A 2 question: O? choice: P A 1 A 2 P C only if O C (A 1 A 2 ) before attack: O i (i = 1 or 2) defense: P A i P/O-symmetry of argumentation forms is given up.

Implications as Rules: Dialogues and Strategies Dialogues (D ) P may assert an atomic formula without O having asserted it before. (E) O can only react on the immediately preceding P-move. (F) O can question a formula A if and only if (i) A has not yet been asserted by O, or (ii) A has already been attacked by P. (Strategies defined as before.) Corresponds to sequent calculus with alternative schema Γ A Γ, A B B Yields dialogical interpretation of implications-as-rules concept.

Implications as Rules: Example 0. P (a b) ((b c) (a c)) 1. O? question 2. P (a b) ((b c) (a c)) choice 3. O a b attack assuming rule b a 4. P (b c) (a c) defense

Implications as Rules: Example 0. P (a b) ((b c) (a c)) 1. O? question 2. P (a b) ((b c) (a c)) choice 3. O a b attack (assuming rule b a) 4. P (b c) (a c) defense 5. O? question 6. P (b c) (a c) choice 7. O b c attack assuming rule c b 8. P a c defense

Implications as Rules: Example 0. P (a b) ((b c) (a c)) 1. O? question 2. P (a b) ((b c) (a c)) choice 3. O a b attack (assuming rule b a) 4. P (b c) (a c) defense 5. O? question 6. P (b c) (a c) choice 7. O b c attack (assuming rule c b) 8. P a c defense 9. O? question 10. P a c choice 11. O a attack 12. P c defense

Implications as Rules: Example 0. P (a b) ((b c) (a c)) 1. O? question 2. P (a b) ((b c) (a c)) choice 3. O a b attack (assuming rule b a) 4. P (b c) (a c) defense 5. O? question 6. P (b c) (a c) choice 7. O b c attack (assuming rule c b) 8. P a c defense 9. O? question 10. P a c choice 11. O a attack 12. P c defense 13. O? question 14. P b choice using rule c b

Implications as Rules: Example 0. P (a b) ((b c) (a c)) 1. O? question 2. P (a b) ((b c) (a c)) choice 3. O a b attack (assuming rule b a) 4. P (b c) (a c) defense 5. O? question 6. P (b c) (a c) choice 7. O b c attack (assuming rule c b) 8. P a c defense 9. O? question 10. P a c choice 11. O a attack 12. P c defense 13. O? question 14. P b choice (using rule c b) 15. O? question 16. P a choice using rule b a

Implications as Rules: Example 0. P (a b) ((b c) (a c)) 1. O? question 2. P (a b) ((b c) (a c)) choice 3. O a b attack (assuming rule b a) 4. P (b c) (a c) defense 5. O? question 6. P (b c) (a c) choice 7. O b c attack (assuming rule c b) 8. P a c defense 9. O? question 10. P a c choice 11. O a attack 12. P c defense 13. O? question 14. P b choice (using rule c b) 15. O? question 16. P a choice (using rule b a) O cannot question P a due to (F): a asserted by O before and not attacked by P. Dialogue is won by P and is a strategy for (a b) ((b c) (a c)).

Implications as Rules and Cut Argumentation form for Cut: assertion: O A (or O?,... ) attack: P B defense: O B 0. P a ((a (b c)) b) 1. O? [0, question] 2. P a ((a (b c)) b) [1, choice] 3. O a [2, attack] 4. P (a (b c)) b [3, defense] 5. O? [4, question] 6. P (a (b c)) b [5, choice] 7. O a (b c) [6, attack] (assuming rule (b c) a) 8. P b c [Cut] 9. O b c [Cut] O? [8, question] 10. P 1 [9, attack] P a [9, choice] (using rule (b c) a) 11. O b [10, defense] 12. P b [7, defense]