conflict directed clause learning for the maximum
play

Conflict Directed Clause Learning for the Maximum Weighted Clique - PowerPoint PPT Presentation

Nov 12, 2022 •147 likes •915 views

Conflict Directed Clause Learning for the Maximum Weighted Clique Problem Emmanuel Hebrard 1 and George Katsirelos 2 1 LAAS-CNRS, Universit 2 MIAT, INRA e de Toulouse The Maximum Weight Clique Problem 5 b d 9 Given a graph G with weight

  1. Conflict Directed Clause Learning for the Maximum Weighted Clique Problem Emmanuel Hebrard 1 and George Katsirelos 2 1 LAAS-CNRS, Universit´ 2 MIAT, INRA e de Toulouse

  2. The Maximum Weight Clique Problem 5 b d 9 Given a graph G with weight function w 10 3 a e c 3 f 13 1 / 16

  3. The Maximum Weight Clique Problem 5 b d 9 Given a graph G with weight function w 10 3 Find a clique – set of vertices all a e pairwise adjacent – of maximum weight c 3 f 13 1 / 16

  4. The Maximum Weight Clique Problem 5 5 b b d 9 Given a graph G with weight function w 10 3 Find a clique – set of vertices all a e pairwise adjacent – of maximum weight c 3 f f 13 13 1 / 16

  5. The Maximum Weight Clique Problem 5 5 b b d 9 Given a graph G with weight function w 10 3 Find a clique – set of vertices all a e pairwise adjacent – of maximum weight c 3 ⇐ ⇒ Find a set of vertices without f f 13 13 edge (Independent Set – IS ) of maximum weight (on the complement) 1 / 16

  6. The Maximum Weight Clique Problem 5 5 b b d 9 Given a graph G with weight function w 10 3 Find a clique – set of vertices all a e pairwise adjacent – of maximum weight c 3 ⇐ ⇒ Find a set of vertices without f f 13 13 edge (Independent Set – IS ) of maximum weight (on the complement) ⇐ ⇒ Find a set of vertices touching every edge (Vertex Cover – VC ) minimum weight (on the complement) 1 / 16

  7. The Maximum Weight Clique Problem 5 b d d 9 9 Given a graph G with weight function w 10 10 3 3 Find a clique – set of vertices all a a e e pairwise adjacent – of maximum weight c c 3 3 ⇐ ⇒ Find a set of vertices without f 13 edge (Independent Set – IS ) of maximum weight (on the complement) ⇐ ⇒ Find a set of vertices touching every edge (Vertex Cover – VC ) minimum weight (on the complement) 1 / 16

  8. The Maximum Weight Clique Problem 5 b d d 9 9 Given a graph G with weight function w 10 10 3 3 Find a clique – set of vertices all a a e e pairwise adjacent – of maximum weight c c 3 3 ⇐ ⇒ Find a set of vertices without f 13 edge (Independent Set – IS ) of maximum weight (on the complement) ⇐ ⇒ Find a set of vertices touching every edge (Vertex Cover – VC ) Error- minimum weight (on the complement) Subgraph correcting Perfect philogeny isomorphism codes 1 / 16

  9. Branch & Bound (CP) vs SAT Encoding Branch & Bound (CP) SAT Encoding 2 / 16

  10. Branch & Bound (CP) vs SAT Encoding Branch & Bound (CP) SAT Encoding Dedicated upper bound (coloring) Weak upper bound (UP) 2 / 16

  11. Branch & Bound (CP) vs SAT Encoding Branch & Bound (CP) SAT Encoding Dedicated upper bound (coloring) Weak upper bound (UP) Dedicated data structure (bitsets) No data structure 2 / 16

  12. Branch & Bound (CP) vs SAT Encoding Branch & Bound (CP) SAT Encoding Dedicated upper bound (coloring) Weak upper bound (UP) Dedicated data structure (bitsets) No data structure Dedicated strategies (degree, color) VSIDS 2 / 16

  13. Branch & Bound (CP) vs SAT Encoding Branch & Bound (CP) SAT Encoding Dedicated upper bound (coloring) Weak upper bound (UP) Dedicated data structure (bitsets) No data structure Dedicated strategies (degree, color) VSIDS Backtracking Conflict-Driven Clause Learning 2 / 16

  14. Branch & Bound (CP) vs SAT Encoding Branch & Bound (CP) SAT Encoding Dedicated upper bound (coloring) Weak upper bound (UP) Dedicated data structure (bitsets) No data structure Dedicated strategies (degree, color) VSIDS Backtracking Conflict-Driven Clause Learning We want the best of both worlds! 2 / 16

  15. Generalised Nogoods [Katsirelos and Bacchus 05][Ohrimenko, Stuckey and Codish 07] Branch and propagate as a CP solver 3 / 16

  16. Generalised Nogoods [Katsirelos and Bacchus 05][Ohrimenko, Stuckey and Codish 07] Branch and propagate as a CP solver Store every deduction made during propagation as an explanation clause ( p 1 ∧ p 2 ∧ . . . ∧ p k ) = ⇒ c with p i ’s and c literals of the form x { = , � = } v 3 / 16

  17. Generalised Nogoods [Katsirelos and Bacchus 05][Ohrimenko, Stuckey and Codish 07] Branch and propagate as a CP solver Store every deduction made during propagation as an explanation clause ( p 1 ∧ p 2 ∧ . . . ∧ p k ) = ⇒ c with p i ’s and c literals of the form x { = , � = } v When failing, compute a conflict using the explanation graph 3 / 16

  18. Generalised Nogoods [Katsirelos and Bacchus 05][Ohrimenko, Stuckey and Codish 07] Branch and propagate as a CP solver Store every deduction made during propagation as an explanation clause ( p 1 ∧ p 2 ∧ . . . ∧ p k ) = ⇒ c with p i ’s and c literals of the form x { = , � = } v When failing, compute a conflict using the explanation graph A global constraint “ X is an independent set of weight larger than k ” ◮ Compute an upper bound ◮ Prune w.r.t. this upper bound ◮ Need to compute explanations! 3 / 16

  19. Branching 5 9 b d For any vertex v : 10 3 a e c 3 f 13 4 / 16

  20. Branching 5 9 b d For any vertex v : 10 10 3 either v is in VC ( x v = true ) a a e c 3 f 13 4 / 16

  21. Branching 5 9 b d For any vertex v : 10 3 either v is in VC ( x v = true ) a e c 3 f 13 4 / 16

  22. Branching 5 9 b d For any vertex v : 10 10 3 either v is in VC ( x v = true ) a a e c 3 or v is in IS ( x v = false ) f 13 4 / 16

  23. Branching 5 5 9 9 b b d d For any vertex v : 10 10 3 either v is in VC ( x v = true ) a a e c c 3 3 or v is in IS ( x v = false ) f f 13 13 4 / 16

  24. Upper bound Unweighted case: coloring 5 b d 9 10 3 a e c 3 f 13 5 / 16

  25. Upper bound Unweighted case: coloring 5 b d 9 A clique of size k requires k colors 10 3 a e c 3 f 13 5 / 16

  26. Upper bound Unweighted case: coloring 5 b d 9 A clique of size k requires k colors 10 3 The chromatic number χ ( G ) of G is an a e upper bound on its clique number ω ( G ) c 3 f 13 5 / 16

  27. Upper bound Unweighted case: coloring 5 5 b b d d 9 9 A clique of size k requires k colors 10 10 3 3 The chromatic number χ ( G ) of G is an a a e e upper bound on its clique number ω ( G ) c c 3 3 We work on the complement, so f f 13 13 non-neighbors cannot share the same color ⇒ clique cover 5 / 16

  28. Upper bound Unweighted case: clique cover An independent set of size k requires k 5 5 b b d d 9 9 cliques 10 10 3 3 The chromatic number χ ( G ) of the a a e e complement of G is an upper bound on c c G ’s independence number α ( G ) 3 3 f f 13 13 We work on the complement, so non-neighbors cannot share the same color ⇒ clique cover 5 / 16

  29. Upper bound Weighted case: clique multi cover [Babel 94] 5 b d 9 10 3 a e c 3 f 13 6 / 16

  30. Upper bound Weighted case: clique multi cover [Babel 94] 5 b d 9 A vertex of weight w must be covered by w cliques 10 3 a e c 3 f 13 6 / 16

  31. Upper bound Weighted case: clique multi cover [Babel 94] 5 b d 9 A vertex of weight w must be covered by w cliques 10 3 a e c The total number of clique is an upper 3 bound on weight of the heaviest independent set f 13 6 / 16

  32. Upper bound Weighted case: clique multi cover [Babel 94] 5 b d 9 A vertex of weight w must be covered by w cliques 10 3 a e c The total number of clique is an upper 3 bound on weight of the heaviest independent set f 13 Babel’s complexity depends on the number of colors ( | E | k ) 6 / 16

  33. Upper bound Weighted case: clique multi cover [Babel 94] 5 b d 9 A vertex of weight w must be covered by w cliques 10 3 a e c The total number of clique is an upper 3 bound on weight of the heaviest independent set f 13 Babel’s complexity depends on the number of colors ( | E | k ) 6 / 16

  34. Upper bound Weighted case: clique multi cover [Babel 94] 5 b d 9 A vertex of weight w must be covered by w cliques 10 3 a e c The total number of clique is an upper 3 bound on weight of the heaviest independent set f 13 Babel’s complexity depends on the number of colors ( | E | k ) 6 / 16

  35. Upper bound Weighted case: clique multi cover [Babel 94] 5 b d 9 A vertex of weight w must be covered by w cliques 10 3 a e c The total number of clique is an upper 3 bound on weight of the heaviest independent set f 13 Babel’s complexity depends on the number of colors ( | E | k ) 6 / 16

  36. Upper bound Weighted case: clique multi cover [Babel 94] 5 5 b b d d 9 9 A vertex of weight w must be covered by w cliques 10 10 3 3 a a e e c c The total number of clique is an upper 3 3 bound on weight of the heaviest independent set f f 13 13 Babel’s complexity depends on the number of colors ( | E | k ) ◮ method similar to [Tavares 16] ( | V | 3 ) 6 / 16

  37. Upper bound Weighted case: clique multi cover [Babel 94] 5 b d 9 A vertex of weight w must be covered by w cliques 10 3 a e c The total number of clique is an upper 3 bound on weight of the heaviest independent set f 13 Babel’s complexity depends on the number of colors ( | E | k ) 3 + 3 ◮ method similar to [Tavares 16] ( | V | 3 ) 6 / 16

Recommend

12/7/2006 Massachusetts Institute of Technology Forward Conflict-Directed Search Backward

12/7/2006 Massachusetts Institute of Technology Forward Conflict-Directed Search Backward

12/7/2006 Massachusetts Institute of Technology Forward Conflict-Directed Search Backward conflict-directed search uses conflicts to select backtrack points and as a cache used to prune nodes. Generalized Conflict Learning for

340 views • 6 slides
Conflict-Driven Clause Learning Marijn J.H. Heule http://www.cs.cmu.edu/~mheule/15816-f19/

Conflict-Driven Clause Learning Marijn J.H. Heule http://www.cs.cmu.edu/~mheule/15816-f19/

Conflict-Driven Clause Learning Marijn J.H. Heule http://www.cs.cmu.edu/~mheule/15816-f19/ Automated Reasoning and Satisfiability, September 16, 2019 1/33 The Satisfiability (SAT) problem ( x 5 x 8 x 2 ) ( x 2 x 1 x

881 views • 74 slides
Conflict-Driven Clause Learning Current Trends in AI Bart Bogaerts March 13, 2020 1 THANKS

Conflict-Driven Clause Learning Current Trends in AI Bart Bogaerts March 13, 2020 1 THANKS

Conflict-Driven Clause Learning Current Trends in AI Bart Bogaerts March 13, 2020 1 THANKS Many slides provided by Joao Marques Silva Material of a SAT/SMT summer school http: //satsmt2013.ics.aalto.fi/slides/Marques-Silva.pdf

1.07k views • 106 slides
Conflict-Driven Clause Learning Armin Biere SAT/SMT/AR Summer School 2019 Lisbon, Portugal July

Conflict-Driven Clause Learning Armin Biere SAT/SMT/AR Summer School 2019 Lisbon, Portugal July

Conflict-Driven Clause Learning Armin Biere SAT/SMT/AR Summer School 2019 Lisbon, Portugal July 4, 2019 Dress Code Summer School Speaker as SAT Problem propositional logic: variables tie shirt negation (not) disjunction (or)

1.09k views • 82 slides
Clause Size Reduction with all-UIP Learning Nick Feng and Fahiem Bacchus University of Toronto

Clause Size Reduction with all-UIP Learning Nick Feng and Fahiem Bacchus University of Toronto

Clause Size Reduction with all-UIP Learning Nick Feng and Fahiem Bacchus University of Toronto Introduction Clause Learning Clause Learning is an essential technique in SAT solvers (1996) u Clause Learning Clause Learning is an essential

664 views • 51 slides
X Schedule 4 Transactional Charges Clause 9 Clause 10 2 RWE npower

X Schedule 4 Transactional Charges Clause 9 Clause 10 2 RWE npower

RWE npower DUoSA Consolidation Group Payment Terms Sub-group 28 th July 2005 General Framework Existing Proposed Clause 6 Charges Clause 6 Clause 7 Billing & Payments Clause 7 Clause 8 Site

390 views • 4 slides
Mini-tutorial on conflict-driven clause learning solvers Marijn J. H. Heule The University of

Mini-tutorial on conflict-driven clause learning solvers Marijn J. H. Heule The University of

Mini-tutorial on conflict-driven clause learning solvers Marijn J. H. Heule The University of Texas at Austin 20/01/2014 @ BIRS Marijn J. H. Heule (UT) Mini-tutorial on CDCL solvers BIRS, January 2014 1 / 24 The Satisfiability (SAT) problem

1.13k views • 59 slides
Maximum Integer Flow in Directed Planar Graphs with Vertex Capacities and Multiple Sources and

Maximum Integer Flow in Directed Planar Graphs with Vertex Capacities and Multiple Sources and

Maximum Integer Flow in Directed Planar Graphs with Vertex Capacities and Multiple Sources and Sinks Yipu Wang University of Illinois at Urbana-Champaign Jan 6, 2019 Maximum flow problem Input: Directed graph G = ( V , E ) source s V ,

437 views • 21 slides
? Results of CASC (2016) Very Sketchy Anatomy of Winning ATPs First/Higher-Order Theorem

? Results of CASC (2016) Very Sketchy Anatomy of Winning ATPs First/Higher-Order Theorem

Proof Search in Conflict Resolution Lifting CDCL (Conflict-Driven Clause Learning) to First-Order Logic Bruno Woltzenlogel Paleo joint work with: Daniyar Itegulov (ITMO, St. Petersburg, Russia) Ezequiel Postan (National University

911 views • 46 slides
On Clause Learning Algorithms for Satisfiability Sam Buss SDF@60 Kurt G odel Research Center

On Clause Learning Algorithms for Satisfiability Sam Buss SDF@60 Kurt G odel Research Center

Clause Learning On Clause Learning Algorithms for Satisfiability Sam Buss SDF@60 Kurt G odel Research Center Vienna July 13, 2013 Clause Learning Satisfiability Satisfiability ( Sat ) An instance of Satisfiability ( Sat ) is a set of

484 views • 23 slides
Proofs in Conflict-Driven Theory Combination Maria Paola Bonacina, Stphane Graham-Lengrand, and

Proofs in Conflict-Driven Theory Combination Maria Paola Bonacina, Stphane Graham-Lengrand, and

Proofs in Conflict-Driven Theory Combination Maria Paola Bonacina, Stphane Graham-Lengrand, and Natarajan Shankar CPP2018, 9th January 2018 1/30 Context: Satisfiability Modulo Theories (SMT) CDCL (Conflict-Driven Clause Learning)

1.57k views • 122 slides
CHAMBERS OF TAX CONSULTANTS Clause by Clause Analysis and Reporting Requirements in Tax Audit

CHAMBERS OF TAX CONSULTANTS Clause by Clause Analysis and Reporting Requirements in Tax Audit

CHAMBERS OF TAX CONSULTANTS Clause by Clause Analysis and Reporting Requirements in Tax Audit Report Presentation by August 18, 2018 Yogesh Thar 1 CONTENTS Clause Number Particulars of the Clause 30A* Impact of Primay adjustments 30B*

1.31k views • 115 slides
Cooperation control in Parallel SAT Solving N. Lazaar, Y. Hamadi, S. Jabbour, M. Sebag Parallel

Cooperation control in Parallel SAT Solving N. Lazaar, Y. Hamadi, S. Jabbour, M. Sebag Parallel

Cooperation control in Parallel SAT Solving N. Lazaar, Y. Hamadi, S. Jabbour, M. Sebag Parallel SAT Solving Decentralized resolution Each core: conflict-directed clause learning Cooperation: each core sends the learned clauses to other

277 views • 11 slides
E ff ectively supporting self - directed learning is one of the critical challenges in

E ff ectively supporting self - directed learning is one of the critical challenges in

E ff ectively supporting self - directed learning is one of the critical challenges in supporting lifelong learning - G. Fischer 1 Do 21st Century Learning Environments Support Self - Directed Learning?: Middle School Students Response

486 views • 33 slides
2020/21 Energy Price Limits Workshop Slides Introduction Under clause 6.20.6 of the WEM Rules,

2020/21 Energy Price Limits Workshop Slides Introduction Under clause 6.20.6 of the WEM Rules,

2020/21 Energy Price Limits Workshop Slides Introduction Under clause 6.20.6 of the WEM Rules, AEMO must annually review the appropriateness of the value of the Maximum STEM Price and the Alternative Maximum STEM Price. Marsden Jacob was

602 views • 26 slides
Look Back Search Cedric Geissmann University Basel October 31, 2014 Outline Backtracking 1

Look Back Search Cedric Geissmann University Basel October 31, 2014 Outline Backtracking 1

Look Back Search Cedric Geissmann University Basel October 31, 2014 Outline Backtracking 1 Backjumping 2 Gaschnigs Backjumping Graph-Based Backjumping Conflict-Directed Backjumping Learning 3 Graph-Based Learning Conflict-Directed

1.08k views • 74 slides
Maximum Flow The Maximum Flow Problem Given Directed graph G = ( V , E ) A capacity

Maximum Flow The Maximum Flow Problem Given Directed graph G = ( V , E ) A capacity

Maximum Flow The Maximum Flow Problem Given Directed graph G = ( V , E ) A capacity function c : E R + Two vertices s and t . s , t -Flow Function f : E R + such that f ( e ) c ( e ) , for all v V

169 views • 12 slides
Lower bounds for width-restricted with Lemmas The Pigeonhole clause learning Principle The

Lower bounds for width-restricted with Lemmas The Pigeonhole clause learning Principle The

Width-restricted clause learning Jan Johannsen Resolution Trees Lower bounds for width-restricted with Lemmas The Pigeonhole clause learning Principle The Ordering Principle Small Width Jan Johannsen Formulas Institut f ur

272 views • 25 slides
E-Learning Day 7 March 2019 To promote self-directed Objectives learning using ICT To ensure

E-Learning Day 7 March 2019 To promote self-directed Objectives learning using ICT To ensure

E-Learning Day 7 March 2019 To promote self-directed Objectives learning using ICT To ensure continuity of learning in the event of unscheduled school closure SLS website Student Learning Space Select the URL below to (SLS) access

459 views • 23 slides
Maximum Entropy Inverse Reinforcement Learning Nomenclature Basis Feature Expectation Matching

Maximum Entropy Inverse Reinforcement Learning Nomenclature Basis Feature Expectation Matching

Maximilian Luz Algorithms for Imitation Learning Summer Semester 2019 MLR/IPVS Maximum Entropy Inverse Reinforcement Learning Nomenclature Basis Feature Expectation Matching Principle of Maximum Entropy Maximum Entropy IRL Algorithm and

480 views • 22 slides
CS480/680 Machine Learning Lecture 6: January 23 st , 2020 Maximum A posteriori & Maximum

CS480/680 Machine Learning Lecture 6: January 23 st , 2020 Maximum A posteriori & Maximum

CS480/680 Machine Learning Lecture 6: January 23 st , 2020 Maximum A posteriori & Maximum Likelihood Zahra Sheikhbahaee Sources: A Tutorial on Energy Based Learning University of Waterloo CS480/680 Winter 2020 Zahra Sheikhbahaee Outline

702 views • 17 slides
EFFICIENT MAXIMUM FLOW ALGORITHM Nikolay Sakharnykh Hugo Braun; May 10, 2017 MAXIMUM FLOW

EFFICIENT MAXIMUM FLOW ALGORITHM Nikolay Sakharnykh Hugo Braun; May 10, 2017 MAXIMUM FLOW

EFFICIENT MAXIMUM FLOW ALGORITHM Nikolay Sakharnykh Hugo Braun; May 10, 2017 MAXIMUM FLOW Definition Example : How much instant power can Palo Alto get using that electric grid? s 30 25 Power SF plant 1 8 20 Directed graph

692 views • 36 slides
A) We would not have missed the train if we had got up earlier. 1. MAIN CLAUSE: 2. DEPENDENT

A) We would not have missed the train if we had got up earlier. 1. MAIN CLAUSE: 2. DEPENDENT

Identifique a MAIN CLAUSE, a DEPENDENT CLAUSE, a combinao verbal e a condio estabelecida pelas IF CLAUSES a seguir: A) We would not have missed the train if we had got up earlier. 1. MAIN CLAUSE: 2. DEPENDENT CLAUSE: 3. COMBINAO

289 views • 8 slides
Optimization Ding Li 1 , Yuchen Jin 1 , Cagri Sahin 2 , James Clause 2 , William G. J. Halfond 1

Optimization Ding Li 1 , Yuchen Jin 1 , Cagri Sahin 2 , James Clause 2 , William G. J. Halfond 1

Integrated Energy-Directed Test Suite Optimization Ding Li 1 , Yuchen Jin 1 , Cagri Sahin 2 , James Clause 2 , William G. J. Halfond 1 1. Department of Computer Science University of Southern California 2. Computer and Information Sciences

799 views • 28 slides

More recommend