Merrick L. Furst

Fast planning through planning graph analysis

We introduce a new approach to planning in STRIPS-like domains based on constructing and analyzing a compact structure we call a Planning Graph. We describe a new planner, Graphplan, that uses this paradigm. Graphplan always returns a shortest-possible partial-order plan, or states that no valid plan exists. We provide empirical evidence in favor of this approach, showing that Graphplan outperforms the total-order planner, Prodigy, and the partial-order planner, UCPOP, on a variety of interesting natural and artificial planning problems. We also give empirical evidence that the plans produced by Graphplan are quite sensible. Since searches made by this approach are fundamentally different from the searches of other common planning methods, they provide a new perspective on the planning problem.

Polynomial-time algorithms for permutation groups

A permutation group on n letters may always be represented by a small set of generators, even though its size may be exponential in n. We show that it is practical to use such a representation since many problems such as membership testing, equality testing, and inclusion testing are decidable in polynomial time. In addition, we demonstrate that the normal closure of a subgroup can be computed in polynomial time, and that this proceaure can be used to test a group for solvability. We also describe an approach to computing the intersection of two groups. The procedures and techniques have wide applicability and have recently been used to improve many graph isomorphism algorithms.

Parity, circuits, and the polynomial-time hierarchy

A super-polynomial lower bound is given for the size of circuits of fixed depth computing the parity function. Introducing the notion of polynomial-size, constant-depth reduction, similar results are shown for the majority, multiplication, and transitive closure functions. Connections are given to the theory of programmable logic arrays and to the relativization of the polynomial-time hierarchy.

Weakly learning DNF and characterizing statistical query learning using Fourier analysis

We present new results, both positive and negative, on the well-studied problem of learning disjunctive normal form (DNF) expressions. We first prove that an algorithm due to Kushilevitz and Mansour [16] can be used to weakly learn DNF using membership queries in polynomial time, with respect to the uniform distribution on the inputs. This is the first positive result for learning unrestricted DNF expressions in polynomial time in any nontrivial formal model of learning. It provides a sharp contrast with the results of Kharitonov [15], who proved that ACO is not efficiently learnable in the same model (given certain plausible cryptographic assumptions). We also present efficient learning algorithms in various models for the read-k and SAT-k subclasses of DNF. For our negative results, we turn our attention to the recently introduced statistical query model of learning [11]. This model is a restricted version of the popular Probably Approximately Correct (PAC) model [23], and practically every class known to be efficiently learnable in the PAC model is in fact learnable in the statistical query model [11]. Here we give a general characterization of the complexity of statistical query learning in terms of the number of uncorrelated functions in the concept class. This is a distributiondependent quantity yielding upper and lower bounds on the number of st atistical queries required for learning on any input distribution. As a corollary, we obtain that DNF expressions and decision trees are not even weakly learnable with ●This research M sponsored in part by the Wr]ght Laboratory, Aeronautical Systems Center, Air Force Materiel Command, USAF, and the Advanced Research Projects Agency (ARPA) under grant number F33615-93-1-1330 Support also M sponsored by the National Sc]ence Foundation under Grant No CC-91 19319. Blum also supported m part by NSF National Young Investigator grant CCR9357793 Views and conclusions contained in this document are those of the authors and should not be interpreted as necessarily representing official po!lcles or endorsements, either expressed or implied, of Wright Laboratory or the United States Government, or NSF tcontact ~“thor Address: AT&T Bell Laboratcmes, Room 2A423, 600 Mountain Avenue, P.O. Box 636, Murray Hill, NJ 07974 Electronic mail. mkearns@research .at t corn ~Thi~ research ~a~ ~“pported in p~~t by The Israel science Foun. datlon administered by The Israel Academy of Sc]ence and Humanities and by a grant of the Israeli Ministry of Science and Technology Permission to co y without fee all or part of this material is granted provided%atthe copies are not madeordistrftrutectfor direct commercial advantage, the ACM copyright notice and the title of the publication and its date appear, and notice is given that copying is by permission of the Association of Computing Machinery. To copy otherwise, or to republish, requires a fee ancf/or specific permission. STOC 945/94 Montreal, Quebec, Canada Q 1994 ACM 0-89791 -663-8/94/0005..

Cryptographic Primitives Based on Hard Learning Problems

3.50 respect to the uniform input distribution in polynomial time in the statistical query model. This result is informationtheoretic and therefore does not rely on any unproven assumptions. It demonstrates that no simple modification of the existing algorithms in the computaticmal learning theory literature for learning various restricted forms of DNF and decision trees from passive random examples (and also several algorithms proposed in the experimental machine learning communities, such as the ID3 algorithm for decision trees [22] and its variants) will solve the general problem. The unifying tool for all of our results is the Fourier analysis of a finite class of boolean functions 011 the hypercube.

Multi-party protocols

Modern cryptography has had considerable impact on the development of computational learning theory. Virtually every intractability result in Valiant’s model [13] (which is representation-independent in the sense that it does not rely on an artificial syntactic restriction on the learning algorithm’s hypotheses) has at its heart a cryptographic construction [4, 9, 1, 10]. In this paper, we give results in the reverse direction by showing how to construct several cryptographic primitives based on certain assumptions on the difficulty of learning. In doing so, we develop further a line of thought introduced by Impagliazzo and Levin [6].

Hierarchy for Imbedding-Distribution Invariants of a Graph

Many different types of inter-process communication have been examined from a complexity point of view [SP, Y]. We study a new model, in which a collection of processes <italic>P<subscrpt>0</subscrpt>, ..., P<subscrpt>k−1</subscrpt></italic> that share information about a set of integers {a<subscrpt>0</subscrpt>, ...,a<subscrpt>k−1</subscrpt>}, communicate to determine a 0-1 predicate of the numbers. In this new model, tremendous sharing of information is allowed, while no single party is given enough information to determine the predicate on its own. Formally, each <italic>P<subscrpt>i</subscrpt></italic> has access to every a<subscrpt>j</subscrpt> except for a<subscrpt>i</subscrpt>. For simplicity, we only allow the parties to communicate as follows.

Genus distributions for two classes of graphs

Most existing papers about graph imbeddings are concerned with the determination of minimum genus, and various others have been devoted to maximum genus or to highly symmetric imbeddings of special graphs. An entirely different viewpoint is now presented in which one seeks distributional information about the huge family of all cellular imbeddings of a graph into all closed surfaces, instead of focusing on just one imbedding or on the existence of imbeddings into just one surface. The distribution of imbeddings admits a hierarchically ordered class of computable invariants, each of which partitions the set of all graphs into much finer subcategories than the subcategories corresponding to minimum genus or to any other single imbedding surface. Quite low in this hierarchy are invariants such as the average genus, taken over all cellular imbeddings, and the average region size, where “region size” means the number of edge traversals required to complete a tour of a region boundary. Further up in the hierarchy is the multiset of duals of a graph. At an intermediate level are the “imbedding polynomials.” The hierarchy is explored, and several specific calculations of the values of some of the invariants are provided. The main results are concerned with the amount of work needed to derive one invariant from another, when possible, and with principles for computing the algebraic effect of adding an edge or of otherwise combining two graphs.

Finding a maximum-genus graph imbedding

Abstract The set of orientable imbeddings of a graph can be partitioned according to the genus of the imbedding surfaces. A genus-respecting breakdown of the number of orientable imbeddings is obtained for every graph in each of two infinite classes. It is proved that the genus distribution of any member of either class is strongly unimodal. These are the first two infinite classes of graphs for which such calculations have been achieved, except for a few classes, such as trees and cycles, whose members have all their cellular orientable imbeddings in the sphere.

Improved learning of AC 0 functions

The computational complexity of constructing the imbeddings of a given graph into surfaces of different genus is not well understood. In this paper, topological methods and a reduction to linear matroid parity are used to develop a polynomial-time algorithm to find a maximum-genus cellular imbedding. This seems to be the first imbedding algorithm for which the running time is not exponential in the genus of the imbedding surface.