Three ∑ P 2 -complete problems in computational learning theory

The consistency problem associated with a concept classC is to determine, given two setsA andB of examples, whether there exists a conceptc inC such that eachx inA is a positive example ofc and eachy inB is a negative example ofc. We explore in this paper the following intuition: for a concept classC, if the membership problem of determining whether a given example is positive for a concept isNP-complete, then the corresponding consistency problem is likely to be _P ² -complete. To support this intuition, we prove that the following three consistency problems for concept classes of patterns, graphs and generalized Boolean formulas, whose membership problems are known to beNP-complete, are _P ² -complete: (a) given two setsA andB of strings, determine whether there exists a patternp such that every string inA is in the languageL(p) and every string inB is not in the languageL(p); (b) given two setsA andB of graphs, determine whether there exists a graphG such that every graph inA is isomorphic to a subgraph ofG and every graph inB is not isomorphic to any subgraph ofG; and (c) given two setsA andB of Boolean formulas, determine whether there exists a 3-CNF Boolean formula such that for every A, is satisfiable and for every B, is not satisfiable. These results suggest that consistendy problems in machine learning are natural candidates for _P ² -complete problems if the corresponding membership problems are known to beNP-complete.In addition, we prove that the corresponding prediction problems for concept classes of polynomial-time nondeterministic Turing machines, nondeterministic Boolean circuits, generalized Boolean formulas, patterns and graphs are prediction-complete for the classR _NP of all concept classes whose membership problems are inNP.     

Searching minimax game trees under memory space constraint

Games such as CHESS, GO and OTHELLO can be represented by minimax game trees. Among various search procedures to solve such game trees,- and SSS^* are perhaps most well known. Although it is proved that SSS^* explores only a subset of the nodes explored by-, - is commonly believed to be faster in real applications, since it requires very little memory space and hence its storage management cost is low. Contrary to this folklore, however, this paper reports, using the OTHELLO game as an example, that SSS^* is much faster than-. It is also demonstrated that SSS^* can be modified to make the required memory space controllable to some extent, while retaining the high efficiency of the original SSS^*.This research was partially supported by the Ministry of Education, Science and Culture of Japan, under a Scientific Grant-in-Aid.     

A Mathematical Analysis of Pānini’s <Emphasis Type="Italic">Śivasūtras</Emphasis>

In Pninis grammar of Sanskrit one finds the ivastras, a table which defines the natural classes of phonological segments in Sanskrit by intervals. We present a formal argument which shows that, using his representation method, Pninis way of ordering the phonological segments to represent the natural classes is optimal. The argument is based on a strictly set-theoretical point of view depending only on the set of natural classes and does not explicitly take into account the phonological features of the segments, which are, however, implicitly given in the way a language clusters its phonological inventory. The key idea is to link the graph of the Hasse-diagram of the set of natural classes closed under intersection to ivastra-style representations of the classes. Moreover, the argument is so general that it allows one to decide for each set of sets whether it can be represented with Pninis method. Actually, Pnini had to modify the set of natural classes to define it by the ivastras (the segment h plays a special role). We show that this modification was necessary and, in fact, the best possible modification. We discuss how every set of classes can be modified in such a way that it can be defined in a ivastra-style representation.¹     

End-to-end communications management using TMN “X” interfaces

The ongoing integration of LANs and WANs to support global communications and businesses and the emergence of integrated broadband communication services has created an increased demand for cooperation between customers, network and service providers to achieve end-to-end service management. Such a cooperation between autonomous authorities, each defining their own administrative management domains, requires the application of an open standardized framework to facilitate and regulate interworking. Such a framework is given by the ITU-T recommendations on TMN, where the so-called X interface is of particular importance for inter-domain management. In this paper, we explain the role of the TMN X interface within an inter-domain TMN architecture supporting end-to-end communications management. We identify the important issues that need to be addressed for the definition and realization of TMN X interfaces and report about our practical experiences with the implementation of TMN X interfaces in the PREPARE project.     

AI & Society and society

This article looks at the broadest implications of public acceptance of AI. A distinction is drawn between conscious belief in a technology, and organic belief where the technology is incorporated into an unconscious world model. The extent to which we feel threatened by AI's apparent denial of spirit is considered, along with a discussion of how people react to this threat. It is proposed that organic acceptance of AI models would lead to a rebirth of popular spiritual concepts as paradoxical as the New Age ideas that have their roots in the theories of physics. Finally the relevance of this speculation is discussed in terms of how it could impinge upon public acceptability of AI technology.     

The Eigenvalue Problem for the 2D Laplacian in ℋ-Matrix Arithmetic and Application to the Heat and Wave Equation

A class of matrices (-matrices) has recently been introduced by Hackbusch for approximating large and fully populated matrices arising from FEM and BEM applications. These matrices are data-sparse and allow approximate matrix operations of almost linear complexity. In the present paper, we choose a special class of -matrices that provides a good approximation to the inverse of the discrete 2D Laplacian. For these 2D -matrices we study the blockwise recursive schemes for block triangular linear systems of equations and the Cholesky and LDL^T factorization in an approximate arithmetic of almost linear complexity. Using the LDL^T factorization we compute eigenpairs of the discrete 2D Laplacian in -matrix arithmetic by means of a so-called simultaneous iteration for computing invariant subspaces of non-Hermitian matrices due to Stewart. We apply the -matrix techniques to approximate the solutions of the high-frequency 2D wave equation for smooth initial data and the 2D heat equation for arbitrary initial data by spectral decomposition of the discrete 2D Laplacian in, up to logarithmic factors, optimal complexity.     

Relativized polynomial hierarchies extending two levels

The existence of relativized polynomial hierarchies extending exactly two levels is shown. More precisely, oraclesX andY are constructed such thatNP(X) ₂ ^{P, X} = ₂ ^{P, X} and ₂ ^{P, Y} ₂ ^{P, Y} = ₂ ^{P, Y} . These results answer a question posed by Baker, Gill, and Solovay ([1]) and generalize their constructions of relativized polynomial hierarchies extending 0 resp. 1 level; i.e.,P(A) = NP(A) resp.P(B) NP(B) = coNP(B) for recursive oraclesA andB.     

Logic Based Abstractions of Real-Time Systems

When verifying concurrent systems described by transition systems, state explosion is one of the most serious problems. If quantitative temporal information (expressed by clock ticks) is considered, state explosion is even more serious. We present a notion of abstraction of transition systems, where the abstraction is driven by the formulae of a quantitative temporal logic, called qu-mu-calculus, defined in the paper. The abstraction is based on a notion of bisimulation equivalence, called , n-equivalence, where is a set of actions and n is a natural number. It is proved that two transition systems are , n-equivalent iff they give the same truth value to all qu-mu-calculus formulae such that the actions occurring in the modal operators are contained in , and with time constraints whose values are less than or equal to n. We present a non-standard (abstract) semantics for a timed process algebra able to produce reduced transition systems for checking formulae. The abstract semantics, parametric with respect to a set of actions and a natural number n, produces a reduced transition system , n-equivalent to the standard one. A transformational method is also defined, by means of which it is possible to syntactically transform a program into a smaller one, still preserving , n-equivalence.     

Integration of parametric and “ad hoc” second order polymorphism in a calculus with subtyping

In this paper we define an extension ofF [CUG92] to which we add functions that dispatch on different terms according to the type they receive as argument. In other words, we enrich the explicit parametric polymorphism ofF by an explicit ad hoc polymorphism (according the classification of [Str67]). We prove that the calculus we obtain, calledF ^& , enjoys the properties of Church-Rosser and Subject Reduction and that its proof system is coherent. We also define a significant subcalculus for which the subtyping is decidable. This extension has not only a logical interest but it is strongly motivated by the foundation of a broadly used programming style: object-oriented programming. The connections betweenF ^& and object-oriented languages are widely stressed, and the modelling byF ^& of some features of the object-oriented style is described, continuing the work of [CGL96].Part of this work has appeared under the title F ^& : integrating parametric and ad hoc second order polymorphism in the 4th International Workshop on Database Programming Languages. New York City, August 1993.The author was supported by grant n. 203.01.56 of the Consiglio Nazionale delle Ricerche-Comitato Nazionale delle Scienze Matematiche to work at LIENS.     

Reasoning about programs in continuation-passing style

Plotkin's _v -calculus for call-by-value programs is weaker than the -calculus for the same programs in continuation-passing style (CPS). To identify the call-by-value axioms that correspond to on CPS terms, we define a new CPS transformation and an inverse mapping, both of which are interesting in their own right. Using the new CPS transformation, we determine the precise language of CPS terms closed under -transformations, as well as the call-by-value axioms that correspond to the so-called administrative -reductions on CPS terms. Using the inverse mapping, we map the remaining and equalities on CPS terms to axioms on call-by-value terms. On the pure (constant free) set of -terms, the resulting set of axioms is equivalent to Moggi's computational -calculus. If the call-by-value language includes the control operatorsabort andcall-with-current-continuation, the axioms are equivalent to an extension of Felleisenet al.'s _v -C-calculus and to the equational subtheory of Talcott's logic IOCC.This article is a revised and extended version of the conference paper with the same title [42]. The technical report of the same title contains additional material.The authors were supported in part by NSF grant CCR 89-17022 and by Texas ATP grant 91-003604014.     

States and the free orthogonality monoid

Let (X, #) be an orthogonality space such that the lattice C(X, #) of closed subsets of (X, #) is orthomodular and let (, ) denote the free orthogonality monoid over (X, #). Let C₀(, ) be the subset of C(, ), consisting of all closures of bounded orthogonal sets. We show that C₀(, ) is a suborthomodular lattice of C(, ) and we provide a necessary and sufficient condition for C₀(, ) to carry a full set of dispersion free states.The work of the second author on this paper was supported by National Science Foundation Grant GP-9005.     

On the Axioms of Scale Space Theory

We consider alternative scale space representations beyond the well-established Gaussian case that satisfy all reasonable axioms. One of these turns out to be subject to a first order pseudo partial differential equation equivalent to the Laplace equation on the upper half plane {(x, s) ^d × | s > 0}. We investigate this so-called Poisson scale space and show that it is indeed a viable alternative to Gaussian scale space. Poisson and Gaussian scale space are related via a one-parameter class of operationally well-defined intermediate representations generated by a fractional power of (minus) the spatial Laplace operator.     

Distributivity in Łℵ0 and Other Sentential Logics

Certain distributivity results for ukasiewicz's infinite-valued logic ₀ are proved axiomatically (for the first time) with the help of the automated reasoning program OTTER. In addition, nondistributivity results are established for a wide variety of positive substructural logics by the use of logical matrices discovered with the automated model-finding programs MACE and MAGIC.     

Gain Scheduled Output Feedback Control Based on LTI Controller Interpolation that Preserves LPV H ∞ Performance

A new approach to the design of a gain scheduled output feedback controller based on linear time-invariant (LTI) controller interpolation that preserves H performance without a varying-parameter rate feedback is proposed. After the controller design is translated into parameterized linear matrix inequalities about parameter matrices based on a parameter-dependent Lyapunov function, the sufficient condition is given for the partition of the varying-parameter set based on the concept of H performance covering. The varying-parameter set is thus partitioned into sufficiently small subsets. After the constant matrices are found for each of the subsets, the required continuous parameter matrices are obtained by using interpolation. The proposed controller overcomes the drawback that the gain scheduled controller may not be found by using the existing gain scheduled linear parameter varying (LPV) controller synthesis. Moreover, the varying-parameter rate feedback is eliminated and the conservatism of the controller design is reduced by means of limiting the upper bound of the varying-parameter rate. Simulation and experiment results prove the effectiveness of the proposed controller.     

Maximum queue size and hashing with lazy deletion

We answer questions about the distribution of the maximum size of queues and data structures as a function of time. The concept of maximum occurs in many issues of resource allocation. We consider several models of growth, including general birth-and-death processes, the M/G/ model, and a non-Markovian process (data structure) for processing plane-sweep information in computational geometry, called hashing with lazy deletion (HwLD). It has been shown that HwLD is optimal in terms of expected time and dynamic space; our results show that it is also optimal in terms of expectedpreallocated space, up to a constant factor.We take two independent and complementary approaches: first, in Section 2, we use a variety of algebraic and analytical techniques to derive exact formulas for the distribution of the maximum queue size in stationary birth-and-death processes and in a nonstationary model related to file histories. The formulas allow numerical evaluation and some asymptotics. In our second approach, in Section 3, we consider the M/G/ model (which includes M/M/ as a special case) and use techniques from the analysis of algorithms to get optimal big-oh bounds on the expected maximum queue size and on the expected maximum amount of storage used by HwLD in excess of the optimal amount. The techniques appear extendible to other models, such as M/M/1.Research was also done while the author was at Princeton University, supported in part by a Procter Fellowship.Research was also done while the author was on sabbatical at INRIA in Rocquencourt, France, and at Ecole Normale Supérieure in Paris, France. Support was provided in part by National Science Foundation Research Grant DCR-84-03613, by an NSF Presidential Young Investigator Award with matching funds from an IBM Faculty Development Award and an AT&T research grant, by a Guggenheim Fellowship, and by the Office of Naval Research and the Defense Advanced Research Projects Agency under Contract N00014-83-K-0146 and ARPA Order 6320, Amendment 1.     

Provably good approximation algorithms for optimal kinodynamic planning: Robots with decoupled dynamics bounds

We consider the following problem: given a robot system, find a minimal-time trajectory that goes from a start state to a goal state while avoiding obstacles by a speed-dependent safety margin and respecting dynamics bounds. In [1] we developed a provably good approximation algorithm for the minimum-time trajectory problem for a robot system with decoupled dynamics bounds (e.g., a point robot in ³). This algorithm differs from previous work in three ways. It is possible (1) to bound the goodness of the approximation by an error term; (2) to bound the computational complexity of our algorithm polynomially; and (3) to express the complexity as a polynomial function of the error term. Hence, given the geometric obstacles, dynamics bounds, and the error term, the algorithm returns a solution that is-close to optimal and requires only a polynomial (in (1/)) amount of time.We extend the results of [1] in two ways. First, we modify it to halve the exponent in the polynomial bounds from 6d to 3d, so that the new algorithm isO(c ^d N 1/)^3d ), whereN is the geometric complexity of the obstacles andc is a robot-dependent constant. Second, the new algorithm finds a trajectory that matches the optimal in time with an factor sacrificed in the obstacle-avoidance safety margin. Similar results hold for polyhedral Cartesian manipulators in polyhedral environments.The new results indicate that an implementation of the algorithm could be reasonable, and a preliminary implementation has been done for the planar case.This paper describes research done at the Computer Science Robotics Laboratory at Cornell University. Support for our robotics research there is provided in part by the National Science Foundation under Grant Nos. IRI-8802390 and IRI-9000532, by a Presidential Young Investigator award, and in part by the Mathematical Sciences Institute, Intel Corporation, and AT&T Bell Laboratories.     

On Steiner minimal trees withLp distance

LetL _p be the plane with the distanced _p (A ₁ ,A ₂ ) = (¦x ₁ –x ₂¦ ^p + ¦y₁ –y ₂¦^p)^/1p wherex _i andy _i are the cartesian coordinates of the pointA _i . LetP be a finite set of points inL _p . We consider Steiner minimal trees onP. It is proved that, for 1 <p < , each Steiner point is of degree exactly three. Define the Steiner ratio _p to be inf{L _s (P)/L _m (P)¦PL _p } whereL _s (P) andL _m (P) are lengths of the Steiner minimal tree and the minimal spanning tree onP, respectively. Hwang showed ₁ = 2/3. Chung and Graham proved ₂ > 0.842. We prove in this paper that {} = 2/3 and (2/2)_{12 p} 3/2 for anyp.This work was supported in part by the National Science Foundation of China and the President Foundation of Academia Sinica.     

A note on dense and nondense families of complexity classes

In a model for a measure of computational complexity, , for a partial recursive functiont, letR _t denote all partial recursive functions having the same domain ast and computable within timet. Let = {R _t |t is recursive} and let = { |_i is actually the running time function of a computation}. and are partially ordered under set-theoretic inclusion. These partial orderings have been extensively investigated by Borodin, Constable and Hopcroft in [3]. In this paper we present a simple uniform proof of some of their results. For example, we give a procedure for easily calculating a model of computational complexity for which is not dense while is dense. In our opinion, our technique is so transparent that it indicates that certain questions of density are not intrinsically interesting for general abstract measures of computational complexity, . (This is not to say that similar questions are necessarily uninteresting for specific models.)Supported by NSF Research Grants GP6120 and GJ27127.     

Deterministic polynomial-time quantum algorithms for Simon’s problem

D. R. Simon stated a problem, so-called Simons problem, whose computational complexity is in the class BQP but not in BPP, where is the function or oracle given in the problem. This result indicates that BPP may be strictly included in its quantum counterpart, BQP. Later, G. Brassard and P. Høyer showed that Simons problem and its extended version can be solved by a deterministic polynomial time quantum algorithm. That is, these problems are in the class EQP. In this paper, we show that Simons problem and its extended version can be deterministically solved in a simpler and more concrete way than that proposed by G. Brassard and P. Høyer.     

Constructing sets of functions which have a givenF-cardinality

The termF-cardinality of (=F-card()) is introduced whereF: ^{n n} is a partial function and is a set of partial functionsf: ^{n n} . TheF-cardinality yields a lower bound for the worst-case complexity of computingF if only functionsf can be evaluated by the underlying abstract automaton without conditional jumps. This complexity bound isindependent from the oracles available for the abstract machine. Thus it is shown that any automaton which can only apply the four basic arithmetic operations needs (n logn) worst-case time to sortn numbers; this result is even true if conditional jumps witharbitrary conditions are possible. The main result of this paper is the following: Given a total functionF: ^{n n} and a natural numberk, it is almost always possible to construct a set such that itsF-cardinality has the valuek; in addition, can be required to be closed under composition of functionsf,g . Moreover, ifF is continuous, then consists of continuous functions.