On the complexity of trial and error

Motivated by certain applications from physics, biochemistry, economics, and computer science in which the objects under investigation are unknown or not directly accessible because of various limitations, we propose a trial-and-error model to examine search problems in which inputs are unknown. More specifically, we consider constraint satisfaction problems ⋀i Ci, where the constraints Ci are hidden, and the goal is to find a solution satisfying all constraints. We can adaptively propose a candidate solution (i.e., trial), and there is a verification oracle that either confirms that it is a valid solution, or returns the index i of a violated constraint (i.e., error), with the exact content of Ci still hidden. We studied the time and trial complexities of a number of natural CSPs, summarized as follows. On one hand, despite the seemingly very little information provided by the oracle, efficient algorithms do exist for Nash, Core, Stable Matching, and SAT problems, whose unknown-input versions are shown to be as hard as the corresponding known-input versions up to a factor of polynomial. The techniques employed vary considerably, including, e.g., order theory and the ellipsoid method with a strong separation oracle. On the other hand, there are problems whose complexities are substantially increased in the unknown-input model. In particular, no time-efficient algorithms exist for Graph Isomorphism and Group Isomorphism (unless PH collapses or P = NP). The proofs use quite nonstandard reductions, in which an efficient simulator is carefully designed to simulate a desirable but computationally unaffordable oracle. Our model investigates the value of input information, and our results demonstrate that the lack of input information can introduce various levels of extra difficulty. The model accommodates a wide range of combinatorial and algebraic structures, and exhibits intimate connections with (and hopefully can also serve as a useful supplement to) certain existing learning and complexity theories.

[1]  Sanjoy Dasgupta,et al.  Coarse sample complexity bounds for active learning , 2005, NIPS.

[2]  D. Fudenberg,et al.  The Theory of Learning in Games , 1998 .

[3]  Paul W. Goldberg,et al.  The Complexity of Computing a Nash Equilibrium , 2009, SIAM J. Comput..

[4]  Michael E. Saks,et al.  Probabilistic Boolean decision trees and the complexity of evaluating game trees , 1986, 27th Annual Symposium on Foundations of Computer Science (sfcs 1986).

[5]  A Indrayan,et al.  Elements of medical research. , 2004, The Indian journal of medical research.

[6]  Jeffery R. Westbrook,et al.  Robot Navigation with Distance Queries , 2000, SIAM J. Comput..

[7]  Daniel S. Hirschberg,et al.  Lower Bounds for the Stable Marriage Problem and its Variants , 1990, SIAM J. Comput..

[8]  Xiaotie Deng,et al.  Exploring an unknown graph , 1990, Proceedings [1990] 31st Annual Symposium on Foundations of Computer Science.

[9]  Margaret J. Robertson,et al.  Design and Analysis of Experiments , 2006, Handbook of statistics.

[10]  Joan Feigenbaum,et al.  Instance-Hiding Proof Systems , 1999 .

[11]  J. Torán,et al.  Solvable Group Isomorphism is (almost) in NP CoNP , 2004 .

[12]  Miklos Santha,et al.  On the Complexity of Trial and Error for Constraint Satisfaction Problems , 2014, ICALP.

[13]  Stathis Zachos,et al.  Does co-NP Have Short Interactive Proofs? , 1987, Inf. Process. Lett..

[14]  Eric van Damme,et al.  Non-Cooperative Games , 2000 .

[15]  Rocco A. Servedio,et al.  Learning DNF in time 2Õ(n1/3) , 2004, J. Comput. Syst. Sci..

[16]  J. V. Vate.  Linear programming brings marital bliss , 1989 .

[17]  Tim Roughgarden,et al.  Algorithmic Game Theory , 2007 .

[18]  Rolf Klein,et al.  The Polygon Exploration Problem , 2001, SIAM J. Comput..

[19]  Richard S. Sutton,et al.  Reinforcement Learning: An Introduction , 1998, IEEE Trans. Neural Networks.

[20]  Andrzej Pelc,et al.  Exploring unknown undirected graphs , 1999, SODA '98.

[21]  Oded Regev,et al.  On lattices, learning with errors, random linear codes, and cryptography , 2005, STOC '05.

[22]  Susanne Albers,et al.  Exploring Unknown Environments with Obstacles , 1999, SODA '99.

[23]  Martin E. Dyer,et al.  Faster random generation of linear extensions , 1999, SODA '98.

[24]  Umesh V. Vazirani,et al.  An Introduction to Computational Learning Theory , 1994 .

[25]  Telikepalli Kavitha.  Linear time algorithms for Abelian group isomorphism and related problems , 2007, J. Comput. Syst. Sci..

[26]  Xiaotie Deng,et al.  Settling the complexity of computing two-player Nash equilibria , 2007, JACM.

[27]  Xiaotie Deng,et al.  How to learn an unknown environment. I: the rectilinear case , 1998, JACM.

[28]  Rocco A. Servedio,et al.  Learning intersections and thresholds of halfspaces , 2002, The 43rd Annual IEEE Symposium on Foundations of Computer Science, 2002. Proceedings..

[29]  Peter Winkler,et al.  Counting linear extensions is #P-complete , 1991, STOC '91.

[30]  Mihalis Yannakakis,et al.  Shortest Paths Without a Map , 1989, Theor. Comput. Sci..

[31]  Joseph Y. Halpern.  Beyond nash equilibrium: solution concepts for the 21st century , 2008, PODC '08.

[32]  Alejandro López-Ortiz,et al.  Generalized Streets Revisited , 1996, ESA.

[33]  Christos H. Papadimitriou,et al.  Algorithms, Games, and the Internet , 2001, ICALP.

[34]  L. S. Shapley,et al.  College Admissions and the Stability of Marriage , 2013, Am. Math. Mon..

[35]  A. Greiner.  MATCHING PEOPLE AND JOBS , 1947 .

[36]  Martin E. Dyer,et al.  A random polynomial-time algorithm for approximating the volume of convex bodies , 1991, JACM.

[37]  Maria-Florina Balcan,et al.  A discriminative model for semi-supervised learning , 2010, J. ACM.

[38]  Brian A. Davey,et al.  An Introduction to Lattices and Order , 1989 .

[39]  David Cohn,et al.  Active Learning , 2010, Encyclopedia of Machine Learning.

[40]  Ning Chen,et al.  Solving Linear Programming with Constraints Unknown , 2013, ICALP.

[41]  Alvin E. Roth,et al.  Two-Sided Matching: A Study in Game-Theoretic Modeling and Analysis , 1990 .

[42]  Kurt Mehlhorn,et al.  Optimal search for rationals , 2003, Inf. Process. Lett..

[43]  Vladimir Vapnik,et al.  Statistical learning theory , 1998 .

[44]  Adam R. Klivans,et al.  Learning DNF in time 2 Õ(n 1/3 ) . , 2001, STOC 2001.

[45]  Burr Settles,et al.  Active Learning Literature Survey , 2009 .

[46]  Sanjeev Arora,et al.  Computational Complexity: A Modern Approach , 2009 .

[47]  Joseph S. B. Mitchell,et al.  Geometric Shortest Paths and Network Optimization , 2000, Handbook of Computational Geometry.

[48]  Michael L. Littman,et al.  Graphical Models for Game Theory , 2001, UAI.

[49]  Jeffrey C. Jackson.  An Efficient Membership-Query Algorithm for Learning DNF with Respect to the Uniform Distribution , 1997, J. Comput. Syst. Sci..

[50]  H. Peyton Young,et al.  Learning by trial and error , 2009, Games Econ. Behav..

[51]  Christos H. Papadimitriou.  Efficient Search for Rationals , 1979, Inf. Process. Lett..

[52]  Susanne Albers,et al.  Exploring Unknown Environments , 2000, SIAM J. Comput..

[53]  G. Dantzig,et al.  COMPLEMENTARY PIVOT THEORY OF MATHEMATICAL PROGRAMMING , 1968 .

[54]  Alejandro López-Ortiz,et al.  Going Home Through an Unknown Street , 1995, WADS.

[55]  Richard S. Sutton,et al.  Introduction to Reinforcement Learning , 1998 .

[56]  Ryan O'Donnell,et al.  Learning DNF from random walks , 2003, 44th Annual IEEE Symposium on Foundations of Computer Science, 2003. Proceedings..

[57]  Oded Goldreich,et al.  Computational complexity: a conceptual perspective , 2008, SIGA.

[58]  Uwe Schöning.  Graph Isomorphism is in the Low Hierarchy , 1988, J. Comput. Syst. Sci..

[59]  Alvin E. Roth.  Deferred acceptance algorithms: history, theory, practice, and open questions , 2008, Int. J. Game Theory.

[60]  Martin Grötschel,et al.  Geometric Methods in Combinatorial Optimization , 1984 .

[61]  Rolf Klein,et al.  An Optimal Competitive Strategy for Walking in Streets , 2004, SIAM J. Comput..

[62]  J. G. Pierce,et al.  Geometric Algorithms and Combinatorial Optimization , 2016 .

[63]  Rocco A. Servedio,et al.  Agnostically learning halfspaces , 2005, 46th Annual IEEE Symposium on Foundations of Computer Science (FOCS'05).

[64]  Leslie G. Valiant,et al.  A theory of the learnable , 1984, STOC '84.

[65]  Maria-Florina Balcan,et al.  The true sample complexity of active learning , 2010, Machine Learning.

[66]  Mona Singh,et al.  Piecemeal Graph Exploration by a Mobile Robot , 1999, Inf. Comput..

[67]  J. Kahn,et al.  Balancing poset extensions , 1984 .

[68]  Graham R. Brightwell,et al.  Balanced pairs in partial orders , 1999, Discret. Math..

[69]  Alexander A. Sherstov,et al.  Cryptographic Hardness for Learning Intersections of Halfspaces , 2006, 2006 47th Annual IEEE Symposium on Foundations of Computer Science (FOCS'06).

[70]  Daniel Lehmann,et al.  Combinatorial auctions with decreasing marginal utilities , 2001, EC '01.

[71]  Baruch Schieber,et al.  Navigating in Unfamiliar Geometric Terrain , 1997, SIAM J. Comput..

[72]  Dana Angluin.  Queries revisited , 2004, Theor. Comput. Sci..

[73]  Thomas G. Dietterich.  What is machine learning? , 2020, Archives of Disease in Childhood.

[74]  Jeff Kahn,et al.  Entropy and sorting , 1992, STOC '92.

[75]  H. Buhrman,et al.  Complexity measures and decision tree complexity: a survey , 2002, Theor. Comput. Sci..

[76]  Lloyd S. Shapley,et al.  On balanced sets and cores , 1967 .