This paper introduces how human languages can be studied in light of recent development of network theories. There are two directions of exploration. One is to study networks existing in the language system. Various lexical networks can be built based on different relationships between words, being semantic or syntactic. Recent studies have shown that these lexical networks exhibit small-world and scale-free features. The other direction of exploration is to study networks of language users (i.e. social networks of people in the linguistic community), and their role in language evolution. Social networks also show small-world and scale-free features, which cannot be captured by random or regular network models. In the past, computational models of language change and language emergence often assume a population to have a random or regular structure, and there has been little discussion how network structures may affect the dynamics. In the second part of the paper, a series of simulation models of diffusion of linguistic innovation are used to illustrate the importance of choosing realistic conditions of population structure for modeling language change. Four types of social networks are compared, which exhibit two categories of diffusion dynamics. While the questions about which type of networks are more appropriate for modeling still remains, we give some preliminary suggestions for choosing the type of social networks for modeling.

Sequential decision theory formally solves the problem of rational agents in uncertain worlds if the true environmental prior probability distribution is known. Solomonoff's theory of universal induction formally solves the problem of sequence prediction for unknown prior distribution. We combine both ideas and get a parameter-free theory of universal Artificial Intelligence. We give strong arguments that the resulting AIXI model is the most intelligent unbiased agent possible. We outline how the AIXI model can formally solve a number of problem classes, including sequence prediction, strategic games, function minimization, reinforcement and supervised learning. The major drawback of the AIXI model is that it is uncomputable. To overcome this problem, we construct a modified algorithm AIXItl that is still effectively more intelligent than any other time t and length l bounded agent. The computation time of AIXItl is of the order t x 2^l. The discussion includes formal definitions of intelligence order relations, the horizon problem and relations of the AIXI theory to other AI approaches.

Because query execution is the most crucial part of Inductive Logic Programming (ILP) algorithms, a lot of effort is invested in developing faster execution mechanisms. These execution mechanisms typically have a low-level implementation, making them hard to debug. Moreover, other factors such as the complexity of the problems handled by ILP algorithms and size of the code base of ILP data mining systems make debugging at this level a very difficult job. In this work, we present the trace-based debugging approach currently used in the development of new execution mechanisms in hipP, the engine underlying the ACE Data Mining system. This debugger uses the delta debugging algorithm to automatically reduce the total time needed to expose bugs in ILP execution, thus making manual debugging step much lighter.

This paper presents a property of propositional theories un der the answer sets semantics (called Equilibrium Logic for this general syntax): any theory can always be reexpress ed as a strongly equivalent disjunctive logic program, possib ly with negation in the head. We provide two different proofs for this result: one involvin g a syntactic transformation, and one that constructs a program starting from the counterm odels of the theory in the intermediate logic of here-and-there.

Structures that are inherent to data of any type can be of import ance, and hierarchical structure is a prime example. In this work we take text corpora and assess the extent of hierarchical structure among words co nstituting the texts. By comprehensively taking context into account we seek to study hierarchical structures in the domain semantics. The data studied in Rammal et al. (1986) and Murtagh (2004) is point pattern data: observational features with their measurements on many coordinate dimensions. Data may be instead presented as time-varyin g signals and in a similar way, related to the findings of Rammal et al. (1986) and 1 Murtagh (2004), we have investigated ultrametric-related prope rties of time series or 1D signals in Murtagh (2005a).

Canonical correlation analysis is a technique to extract common features from a pair of multivariate data. In complex situations, however, it does not extract useful features because of its linearity. On the other hand, kernel method used in support vector machine is an efficient approach to improve such a linear method. In this paper, we investigate the effectiveness of applying kernel method to canonical correlation analysis.

Correlated time series are time series that, by virtue of the underlying process to which they refer, are expected to influence each other strongly. We introduce a novel approach to handle such time series, one that models their interaction as a two-dimensional cellular automaton and therefore allows them to be treated as a single entity. We apply our approach to the problems of filling gaps and predicting values in rainfall time series. Computational results show that the new approach compares favorably to Kalman smoothing and filtering.

Open answer set programming (OASP) is an extension of answer set programming where one may ground a program with an arbitrary superset of the program's constants. We define a fixed point logic (FPL) extension of Clark's completion such that open answer sets correspond to models of FPL formulas and identify a syntactic subclass of programs, called (loosely) guarded programs. Whereas reasoning with general programs in OASP is undecidable, the FPL translation of (loosely) guarded programs falls in the decidable (loosely) guarded fixed point logic (mu(L)GF). Moreover, we reduce normal closed ASP to loosely guarded OASP, enabling for the first time, a characterization of an answer set semantics by muLGF formulas. We further extend the open answer set semantics for programs with generalized literals. Such generalized programs (gPs) have interesting properties, e.g., the ability to express infinity axioms. We restrict the syntax of gPs such that both rules and generalized literals are guarded. Via a translation to guarded fixed point logic, we deduce 2-exptime-completeness of satisfiability checking in such guarded gPs (GgPs). Bound GgPs are restricted GgPs with exptime-complete satisfiability checking, but still sufficiently expressive to optimally simulate computation tree logic (CTL). We translate Datalog lite programs to GgPs, establishing equivalence of GgPs under an open answer set semantics, alternation-free muGF, and Datalog lite.

Constraint Programming (CP)[1] has been successfully appl ied to both constraint satisfaction and constraint optimization prob lems. A wide variety of specialized global constraints provide critical assistan ce in achieving a good model that can take advantage of the structure of the problem in the search for a solution. However, a key outstanding issue is the representation of'a d-hoc' constraints that do not have an inherent combinatorial nature, and hence are n ot modelled well using narrowly specialized global constraints. We attempt to address this issue by considering a hybrid of search and compilation. Specificall y we suggest the use of Reduced Ordered Multi-V alued Decision Diagrams (ROMDDs) as the supporting data structure for a generic global constraint. We g ive an algorithm for maintaining generalized arc consistency (GAC) on this cons traint that amortizes the cost of the GAC computation over a root-to-leaf path in th e search tree without requiring asymptotically more space than used for the MD D. Furthermore we present an approach for incrementally maintaining the redu ced property of the MDD during the search, and show how this can be used for provid ing domain entailment detection. Finally we discuss how to apply our ap proach to other similar data structures such as AOMDDs and Case DAGs. The techni que used can be seen as an extension of the GAC algorithm for the regular la nguage constraint on finite length input [2].

Speech recognition based on the syllable segment is discussed in this paper. The principal search methods in space of states for the speech recognition problem by segment-syllabic parameters trajectory synthesis are investigated. Recognition as comparison the parameters trajectories in chosen speech units on the sections of the segmented speech is realized. Some experimental results are given and discussed.