Showing papers presented at "Computer Science Logic in 2007"

On acyclic conjunctive queries and constant delay enumeration

Abstract: We study the enumeration complexity of the natural extension of acyclic conjunctive queries with disequalities. In this language, a number of NP-complete problems can be expressed. We first improve a previous result of Papadimitriou and Yannakakis by proving that such queries can be computed in time c.|M|ċ|ϕ(M)| where M is the structure, ϕ(M) is the result set of the query and c is a simple exponential in the size of the formula ϕ. A consequence of our method is that, in the general case, tuples of such queries can be enumerated with a linear delay between two tuples. We then introduce a large subclass of acyclic formulas called CCQ≠ and prove that the tuples of a CCQ≠ query can be enumerated with a linear time precomputation and a constant delay between consecutive solutions. Moreover, under the hypothesis that the multiplication of two n×n boolean matrices cannot be done in time O(n2), this leads to the following dichotomy for acyclic queries: either such a query is in CCQ≠ or it cannot be enumerated with linear precomputation and constant delay. Furthermore we prove that testing whether an acyclic formula is in CCQ≠ can be performed in polynomial time. Finally, the notion of free-connex treewidth of a structure is defined. We show that for each query of free-connex treewidth bounded by some constant k, enumeration of results can be done with O(|M|k+1) precomputation steps and constant delay.

Integrating linear arithmetic into superposition calculus

Abstract: We present a method of integrating linear rational arithmetic into superposition calculus for first-order logic. One of our main results is completeness of the resulting calculus under some finiteness assumptions.

From proofs to focused proofs: a modular proof of focalization in linear logic

Abstract: Probably the most significant result concerning cut-free sequent calculus proofs in linear logic is the completeness of focused proofs. This completeness theorem has a number of proof theoretic applications -- e.g. in game semantics, Ludics, and proof search -- and more computer science applications -- e.g. logic programming, call-by-name/value evaluation. Andreoli proved this theorem for first-order linear logic 15 years ago. In the present paper, we give a new proof of the completeness of focused proofs in terms of proof transformation. The proof of this theorem is simple and modular: it is first proved for MALL and then is extended to full linear logic. Given its modular structure, we show how the proof can be extended to larger systems, such as logics with induction. Our analysis of focused proofs will employ a proof transformation method that leads us to study how focusing and cut elimination interact. A key component of our proof is the construction of a focalization graph which provides an abstraction over how focusing can be organized within a given cut-free proof. Using this graph abstraction allows us to provide a detailed study of atomic bias assignment in a way more refined that is given in Andreoli's original proof. Permitting more flexible assignment of bias will allow this completeness theorem to help establish the completeness of a number of other automated deduction procedures. Focalization graphs can be used to justify the introduction of an inference rule for multifocus derivation: a rule that should help us better understand the relations between sequentiality and concurrency in linear logic.

Not enough points is enough

Abstract: Models of the untyped λ-calculus may be defined either as applicative structures satisfying a bunch of first-order axioms (λ-models), or as reflexive objects in cartesian closed categories (categorical models). In this paper we show that any categorical model of λ-calculus can be presented as a λ-model, even when the underlying category does not have enough points. We provide an example of an extensional model of λ-calculus in a category of sets and relations which has not enough points. Finally, we present some of its algebraic properties which make it suitable for dealing with non-deterministic extensions of λ-calculus.

Focusing and polarization in intuitionistic logic

Abstract: A focused proof system provides a normal form to cut-free proofs that structures the application of invertible and noninvertible inference rules The focused proof system of Andreoli for linear logic has been applied to both the proof search and the proof normalization approaches to computation Various proof systems in literature exhibit characteristics of focusing to one degree or another We present a new, focused proof system for intuitionistic logic, called LJF, and show how other proof systems can be mapped into the new system by inserting logical connectives that prematurely stop focusing We also use LJF to design a focused proof system for classical logic Our approach to the design and analysis of these systems is based on the completeness of focusing in linear logic and on the notion of polarity that appears in Girard's LC and LU proof systems

Precise relational invariants through strategy iteration

Abstract: We present a practical algorithm for computing exact least solutions of systems of equations over the rationals with addition, multiplication with positive constants, minimum and maximum. The algorithm is based on strategy improvement combined with solving linear programming problems for each selected strategy. We apply our technique to compute the abstract least fixpoint semantics of affine programs over the relational template constraint matrix domain [20]. In particular, we thus obtain practical algorithms for computing the abstract least fixpoint semantics over the zone and octagon abstract domain.

The theory of calculi with explicit substitutions revisited

Abstract: Calculi with explicit substitutions (ES) are widely used in different areas of computer science. Complex systems with ES were developed these last 15 years to capture the good computational behaviour of the original systems (with meta-level substitutions) they were implementing. In this paper we first survey previous work in the domain by pointing out the motivations and challenges that guided the development of such calculi. Then we use very simple technology to establish a general theory of explicit substitutions for the lambda-calculus which enjoys fundamental properties such as simulation of one-step beta-reduction, confluence on metaterms, preservation of beta-strong normalisation, strong normalisation of typed terms and full composition. The calculus also admits a natural translation into Linear Logic's proof-nets.

Clique-width and parity games

Abstract: The question of the exact complexity of solving parity games is one of the major open problems in system verification, as it is equivalent to the problem of model-checking the modal µ-calculus. The known upper bound is NP∩co-NP, but no polynomial algorithm is known. It was shown that on tree-like graphs (of bounded tree-width and DAG-width) a polynomial-time algorithm does exist. Here we present a polynomial-time algorithm for parity games on graphs of bounded clique-width (class of graphs containing e.g. complete bipartite graphs and cliques), thus completing the picture. This also extends the tree-width result, as graphs of bounded tree-width are a subclass of graphs of bounded clique-width. The algorithm works in a different way to the tree-width case and relies heavily on an interesting structural property of parity games.

Typed normal form bisimulation

Abstract: Normal form bisimulation is a powerful theory of program equivalence, originally developed to characterize Levy-Longo tree equivalence and Boehm tree equivalence. It has been adapted to a range of un-typed, higher-order calculi, but types have presented a difficulty. In this paper, we present an account of normal form bisimulation for types, including recursive types. We develop our theory for a continuation-passing style calculus, Jump-With-Argument (JWA), where normal form bisimilarity takes a very simple form. We give a novel congruence proof, based on insights from game semantics. A notable feature is the seamless treatment of eta-expansion. We demonstrate the normal form bisimulation proof principle by using it to establish a syntactic minimal invariance result and the uniqueness of the fixed point operator at each type.

A soft type assignment system for &lambda-calculus

Abstract: Soft Linear Logic (SLL) is a subsystem of second-order linear logic with restricted rules for exponentials, which is correct and complete for PTIME. We design a type assignment system for the λ-calculus (STA), which assigns to λ-terms as types (a proper subset of) SLL formulas, in such a way that typable terms inherit the good complexity properties of the logical system. Namely STA enjoys subject reduction and normalization, and it is correct and complete for PTIME and FPTIME.

Classical and intuitionistic logic are asymptotically identical

Abstract: This paper considers logical formulas built on the single binary connector of implication and a finite number of variables. When the number of variables becomes large, we prove the following quantitative results: asymptotically, all classical tautologies are simple tautologies. It follows that asymptotically, all classical tautologies are intuitionistic.

MSO on the infinite binary tree: choice and order

Abstract: We give a new proof showing that it is not possible to define in monadic second-order logic (MSO) a choice function on the infinite binary tree. This result was first obtained by Gurevich and Shelah using set theoretical arguments. Our proof is much simpler and only uses basic tools from automata theory. We discuss some applications of the result concerning unambiguous tree automata and definability of winning strategies in infinite games. In a second part we strengthen the result of the non-existence of an MSO-definable well-founded order on the infinite binary tree by showing that every infinite binary tree with a well-founded order has an undecidable MSO-theory.

A Soft Type Assignment System for λ-Calculus

Abstract: Soft Linear Logic (SLL) is a subsystem of second-order linear logic with restricted rules for exponentials, which is correct and complete for PTIME. We design a type assignment system for the λ-calculus (STA), which assigns to λ-terms as types (a proper subset of) SLL formulas, in such a way that typable terms inherit the good complexity properties of the logical system. Namely STA enjoys subject reduction and normalization, and it is correct and complete for PTIME and FPTIME.

Incorporating tables into proofs

Abstract: We consider the problem of automating and checking the use of previously proved lemmas in the proof of some main theorem. In particular, we call the collection of such previously proved results a table and use a partial order on the table's entries to denote the (provability) dependency relationship between tabled items. Tables can be used in automated deduction to store previously proved subgoals and in interactive theorem proving to store a sequence of lemmas introduced by a user to direct the proof system towards some final theorem. Tables of literals can be incorporated into sequent calculus proofs using two ideas. First, cuts are used to incorporate tabled items into a proof: one premise of the cut requires a proof of the lemma and the other branch of the cut inserts the lemma into the set of assumptions. Second, to ensure that lemma is not reproved, we exploit the fact that in focused proofs, atoms can have different polarity. Using these ideas, simple logic engines that do focused proof search (such as logic programming interpreters) are able to check proofs for correctness with guarantees that previous work is not redone. We also discuss how a table can be seen as a proof object and discuss some possible uses of tables-as-proofs.

Continuous previsions

Abstract: We define strong monads of continuous (lower, upper) previsions, and of forks, modeling both probabilistic and non-deterministic choice. This is an elegant alternative to recent proposals by Mislove, Tix, Keimel, and Plotkin. We show that our monads are sound and complete, in the sense that they model exactly the interaction between probabilistic and (demonic, angelic, chaotic) choice.

Qualitative temporal and spatial reasoning revisited

Abstract: Establishing local consistency is one of the main algorithmic techniques in temporal and spatial reasoning. In this area, one of the central questions for the various proposed temporal and spatial constraint languages is whether local consistency implies global consistency. Showing that a constraint language G has this "local-to-global" property implies polynomial-time tractability of the constraint language, and has further pleasant algorithmic consequences. In the present paper, we study the "local-to-global" property by making use of a recently established connection of this property with universal algebra. Specifically, the connection shows that this property is equivalent to the presence of a so-called quasi near-unanimity polymorphism of the constraint language. We obtain new algorithmic results and give very concise proofs of previously known theorems. Our results concern well-known and heavily studied formalisms such as the point algebra and its extensions, Allen's interval algebra, and the spatial reasoning language RCC-5.

Satisfiability of a spatial logic with tree variables

Abstract: We investigate in this paper the spatial logic TQL for querying semistructured data, represented as unranked ordered trees over an infinite alphabet. This logic consists of usual Boolean connectives, spatial connectives (derived from the constructors of a tree algebra), tree variables and a fixpoint operator for recursion. Motivated by XML-oriented tasks, we investigate the guarded TQL fragment. We prove that for closed formulas this fragment is MSO-complete. In presence of tree variables, this fragment is strictly more expressive than MSO as it allows for tree (dis)equality tests, i.e. testing whether two subtrees are isomorphic or not. We devise a new class of tree automata, called TAGED, which extends tree automata with global equality and disequality constraints. We show that the satisfiability problem for guarded TQL formulas reduces to emptiness of TAGED. Then, we focus on bounded TQL formulas: intuitively, a formula is bounded if for any tree, the number of its positions where a subtree is captured by a variable is bounded. We prove this fragment to correspond with a subclass of TAGED, called bounded TAGED, for which we prove emptiness to be decidable. This implies the decidability of the bounded guarded TQL fragment. Finally, we compare bounded TAGED to a fragment of MSO extended with subtree isomorphism tests.

A cut-free and invariant-free sequent calculus for PLTL

Abstract: Sequent calculi usually provide a general deductive setting that uniformly embeds other proof-theoretical approaches, such as tableaux methods, resolution techniques, goal-directed proofs, etc. Unfortunately, in temporal logic, existing sequent calculi make use of a kind of inference rules that prevent the effective mechanization of temporal deduction in the general setting. In particular, temporal sequent calculi either need some form of cut, or they make use of invariants, or they include infinitary rules. This is the case even for the simplest kind of temporal logic, propositional linear temporal logic (PLTL). In this paper, we provide a complete finitary sequent calculus for PLTL, called FC, that not only is cut-free but also invariant-free. In particular, we introduce new rules which provide a new style of temporal deduction. We give a detailed proof of completeness.

Building decision procedures in the calculus of inductive constructions

Abstract: It is commonly agreed that the success of future proof assistants will rely on their ability to incorporate computations within deduction in order to mimic the mathematician when replacing the proof of a proposition P by the proof of an equivalent proposition P' obtained from P thanks to possibly complex calculations. In this paper, we investigate a new version of the calculus of inductive constructions which incorporates arbitrary decision procedures into deduction via the conversion rule of the calculus. The novelty of the problem in the context of the calculus of inductive constructions lies in the fact that the computation mechanism varies along proof-checking: goals are sent to the decision procedure together with the set of user hypotheses available from the current context. Our main result shows that this extension of the calculus of constructions does not compromise its main properties: confluence, subject reduction, strong normalization and consistency are all preserved.

Classical program extraction in the calculus of constructions

Abstract: We show how to extract classical programs expressed in Krivine λc-calculus from proof-terms built in a proof-irrelevant and classical version of the calculus of constructions with universes. For that, we extend Krivine's realisability model of classical second-order arithmetic to the calculus of constructions with universes using a structure of Π-set which is reminiscent of ω-sets, and show that our realisability model validates Peirce's law and proof-irrelevance. Finally, we extend the extraction scheme to a primitive data-type of natural numbers in a way which preserves the whole compatibility with the classical realisability interpretation of second-order arithmetic.

The Power of Counting Logics on Restricted Classes of Finite Structures

Abstract: Although Cai, Furer and Immerman have shown that fixed-point logic with counting (IFP+C) does not express all polynomial-time properties of finite structures, there have been a number of results demonstrating that the logic does capture P on specific classes of structures. Grohe and Marino showed that IFP+C captures P on classes of structures of bounded treewidth, and Grohe showed that IFP+C captures P on planar graphs. We show that the first of these results is optimal in two senses. We show that on the class of graphs defined by a non-constant bound on the tree-width of the graph, IFP+C fails to capture P. We also show that on the class of graphs whose local tree-width is bounded by a non-constant function, IFP+C fails to capture P. Both these results are obtained by an analysis of the Cai---Furer---Immerman (CFI) construction in terms of the treewidth of graphs, and cops and robber games; we present some other implications of this analysis. We then demonstrate the limits of this method by showing that the CFI construction cannot be used to show that IFP+C fails to capture P on proper minor-closed classes.

Logical refinements of Church's problem

Abstract: Church's Problem (1962) asks for the construction of a procedure which, given a logical specification ϕ on sequence pairs, realizes for any input sequence X an output sequence Y such that (X,Y) satisfies ϕ. Buchi and Landweber (1969) gave a solution for MSO specifications in terms of finite-state automata. We address the problem in a more general logical setting where not only the specification but also the solution is presented in a logical system. Extending the result of Buchi and Landweber, we present several logics L such that Church's Problem with respect to L has also a solution in L, and we discuss some perspectives of this approach.

Structure theorem and strict alternation hierarchy for FO 2 on words

Abstract: It is well-known that every first-order property on words is expressible using at most three variables. The subclass of properties expressible with only two variables is also quite interesting and well-studied. We prove precise structure theorems that characterize the exact expressive power of first-order logic with two variables on words. Our results apply to FO2[<] and FO2[<, Suc], the latter of which includes the binary successor relation in addition to the linear ordering on string positions. For both languages, our structure theorems show exactly whatis expressible using a given quantifier depth, n, and using m blocks of alternating quantifiers, for any m ≤ n. Using these characterizations, we prove, among other results, that there is a strict hierarchy of alternating quantifiers for both languages. The question whether there was such a hierarchy had been completely open. As another consequence of our structural results, we show that satisfiability for FO2[<], which is NEXP-complete in general, becomes NP-complete once we only consider alphabets of a bounded size.

Unbounded proof-length speed-up in deduction modulo

Abstract: In 1973, Parikh proved a speed-up theorem conjectured by Godel 37 years before: there exist arithmetical formulae that are provable in first order arithmetic, but whose shorter proof in second order arithmetic is arbitrarily smaller than any proof in first order On the other hand, resolution for higher order logic can be simulated step by step in a first order narrowing and resolution method based on deduction modulo, whose paradigm is to separate deduction and computation to make proofs clearer and shorter We prove that i+1-th order arithmetic can be linearly simulated into i-th order arithmetic modulo some confluent and terminating rewrite system We also show that there exists a speed-up between i-th order arithmetic modulo this system and i-th order arithmetic without modulo All this allows us to prove that the speed-up conjectured by Godel does not come from the deductive part of the proofs, but can be expressed as simple computation, therefore justifying the use of deduction modulo as an efficient first order setting simulating higher order

Lambda theories of effective lambda models

Abstract: A longstanding open problem is whether there exists a nonsyntactical model of the untyped λ-calculus whose theory is exactly the least λ-theory λβ. In this paper we investigate the more general question of whether the equational/order theory of a model of the untyped λ-calculus can be recursively enumerable (r.e. for brevity). We introduce a notion of effective model of λ-calculus, which covers in particular all the models individually introduced in the literature. We prove that the order theory of an effective model is never r.e.; from this it follows that its equational theory cannot be λβ, λβ. We then show that no effective model living in the stable or strongly stable semantics has an r.e. equational theory. Concerning Scott's semantics, we investigate the class of graph models and prove that no order theory of a graph model can be r.e., and that there exists an effective graph model whose equational/ order theory is the minimum one. Finally, we show that the class of graph models enjoys a kind of downwards Lowenheim-Skolem theorem.

There Exist Some ω-Powers of Any Borel Rank

Abstract: The operation V ?V ? is a fundamental operation over finitary languages leading to ?-languages. Since the set Σ ? of infinite words over a finite alphabet Σ can be equipped with the usual Cantor topology, the question of the topological complexity of ?-powers of finitary languages naturally arises and has been posed by Niwinski [Niw90], Simonnet [Sim92] and Staiger [Sta97a] . It has been recently proved that for each integer n ? 1, there exist some ?-powers of context free languages which are ${\bf \Pi}_n^0$-complete Borel sets,[Fin01], that there exists a context free language L such that L ? is analytic but not Borel,[Fin03] , and that there exists a finitary language V such that V ? is a Borel set of infinite rank, [Fin04]. But it was still unknown which could be the possible infinite Borel ranks of ?-powers. We fill this gap here, proving the following very surprising result which shows that ?-powers exhibit a great topological complexity: for each non-null countable ordinal ?, there exist some -complete ?-powers, and some -complete ?-powers.

Linear realizability

Abstract: We define a notion of relational linear combinatory algebra (rLCA) which is a generalization of a linear combinatory algebra defined by Abramsky, Haghverdi and Scott. We also define a category of assemblies as well as a category of modest sets which are realized by rLCA. This is a linear style of realizability in a way that duplicating and discarding of realizers is allowed in a controlled way. Both categories form linear-non-linear models and their coKleisli categories have a natural number object. We construct some examples of rLCA's which have some relations to well known PCA's.

Subexponential time and fixed-parameter tractability: exploiting the miniaturization mapping

Abstract: Recently a mapping M, the so-called miniaturization mapping, has been introduced and it has been shown that M faithfully translates subexponential parameterized complexity into (unbounded) parameterized complexity [2]. We determine the preimages under M of various (classes of) problems and show that they coincide with natural reparameterizations which take into account the amount of nondeterminism needed to solve them.

Forest expressions

Abstract: We define regular expressions for unranked trees (actually, ordered sequences of unranked trees, called forests). These are compared to existing regular expressions for trees. On the negative side, our expressions have complementation, and do not define all regular languages. On the positive side, our expressions do not use variables, and have a syntax very similar to that of regular expressions for word languages. We examine the expressive power of these expressions, especially from a logical point of view. The class of languages defined corresponds to a form of chain logic [5,6]. Furthermore, the star-free expressions coincide with first-order logic. Finally, we show that a concatenation hierarchy inside the expressions corresponds to the quantifier prefix hierarchy for first-order logic, generalizing a result of Thomas.

Game characterizations and the PSPACE-completeness of tree resolution space

Abstract: The Prover/Delayer game is a combinatorial game that can be used to prove upper and lower bounds on the size of Tree Resolution proofs, and also perfectly characterizes the space needed to compute them. As a proof system, Tree Resolution forms the underpinnings of all DPLL-based SAT solvers, so it is of interest not only to proof complexity researchers, but also to those in the area of propositional reasoning. In this paper, we prove the PSPACE-Completeness of the Prover/Delayer game as well as the problem of predicting Tree Resolution space requirements, where space is the number of clauses that must be kept in memory simultaneously during the computation of a refutation. Since in practice memory is often a limiting resource, researchers developing SAT solvers may wish to know ahead of time how much memory will be required for solving a certain formula, but the present result shows that predicting this is at least as hard as it would be to simply find a refutation.