Showing papers in "Ai Communications in 2006"

The state of CASC

Abstract: The CADE ATP System Competition (CASC) is an annual evaluation of fully automatic, first order Automated Theorem Proving systems - the world championship for such systems. This paper captures the state of CASC after CASC-20, the tenth CASC, held in 2005. It provides a summarized history of CASC, details of the current design of the competition, observations and discussion of the effects of CASC on ATP, lessons learnt during CASC, and remarks regarding the past, present, and future of CASC.

Integrating heterogeneous adaptation techniques to build a flexible and usable mobile tourist guide

Abstract: Intelligent adaptation is a key issue for the design of flexible support systems for mobile users. In this paper we present UbiquiTO, a tourist guide which integrates different forms of adaptation: (i) to the device used (web access via Laptop, PDA, smartphone), (ii) to the user and her features and preferences (personalized interaction), (iii) to the context of interaction, and in particular to the user location, besides some other features such as the time of the day. UbiquiTO adapts the content of the service being provided (recommendation and amount/type of information/features associated with each recommendation) and the presentation (interface). In order to achieve better performance it keeps track of the user behavior, updating and refining the user model during the interaction. In the paper we introduce the architecture of the system and the choices we made as regards user, device and context modeling and adaptation strategies. We also present the results of a preliminary evaluation of the system behavior.

On belief change in ontology evolution: Thesis

Abstract: We study the feasibility of applying the dominating paradigm of belief revision, the AGM theory, to logics outside its original scope, with special emphasis given on logics used for ontological representation in the Semantic Web. Such an application would allow determining whether a proposed change operator (e.g., an ontology evolution algorithm) behaves rationally or not. We give a number of interesting theoretical results related to the (generalized) AGM theory and show that our work can find important applications in dynamic environments employing non-classical logical formalisms. We focus on a particular such environment, namely ontology evolution in the Semantic Web, and show how our work can be used to provide ontology evolution researchers with powerful formal tools from the research area of belief revision.

Evolution of a supply chain management game for the trading agent competition

Abstract: TAC SCM is a supply chain management game for the Trading Agent Competition (TAC). The purpose of TAC is to spur high quality research into realistic trading agent problems. We discuss TAC and TAC SCM: game and competition design, scientific impact, and lessons learnt.With a conscious effort to these ends, a competition may help focus research on well-chosen problems in high impact domains, facilitate comparison of results, and create a community eager to analyze and build on the results of peers and in working jointly towards the perfection of solutions.Potential pitfalls include emphasizing winning to the exclusion of testing scientific hypotheses, and carrying the overhead of designing, implementing, and operating a competition.

Thesis: automatic ontology generation from web tabular structures

Abstract: Turning the current Web into a Semantic Web requires automatic approaches for document annotation, since manual approaches will not scale in general. The focus of the thesis is on automatic transformation of arbitrary table-like structures into knowledge models, i.e., ontologies. The presented work is based on Hurst's table model and consists of a methodology, an accompanying implementation named TARTAR, and a thorough evaluation. The evaluation showed over 80% success rate of automatic transformation of tables into semantic representations and 100% accuracy in the task of query answering over the table contents.

A backjumping technique for disjunctive logic programming

Abstract: In this work we present a backjumping technique for Disjunctive Logic Programming under the Stable Model Semantics (SDLP). It builds upon related techniques that had originally been introduced for constraint solving, which have been adapted to propositional satisfiability testing, and recently also to non-disjunctive logic programming under the stable model semantics (SLP) [1,2].We focus on backjumping without clause learning, providing a new theoretical framework for backjumping in SDLP, elaborating on and exploiting peculiarities of the disjunctive setting. We present a reason calculus and associated computations, which - compared to the traditional approaches - reduces the information to be stored, while fully preserving the correctness and the efficiency of the backjumping technique, handling specific aspects of disjunction in a benign way. We implemented the proposed technique in DLV, the state-of-the-art SDLP system.We have conducted several experiments on hard random and structured instances in order to assess the impact of backjumping. To this end, we have compared DLV in various versions: With and without the backjumping method described in this paper, in combination with two different heuristic functions. Our conclusion is that under any of the heuristic functions, DLV with backjumping is favourable to DLV without backjumping. DLV with backjumping performs particularly well on structured satisfiability and quantified boolean formula instances, where the search space and execution time are effectively cut.

Exploiting agents in e-learning and skills management context

Abstract: Nowadays, it is quite agreed that organizations gain limited advantages in adopting e-learning platforms that only provide educational contents. An advantageous e-learning platform should have instead the capability to help enrich, share and circulate organization knowledge, thus contributing to making the organization dynamic and flexible. In this paper MASEL, a Multi-Agent System for E-Learning and Skill Management is described. MASEL performs the following tasks: (i) supports Chief Learning Officers in defining roles, associated competencies and required knowledge level; (ii) manages the skill map of the organization; (iii) evaluates human resources competence gaps; (iv) supports employees in filling the competence gaps related to their roles; (v) creates personalized learning paths according to feedbacks that users provide to optimize the acquisition of required competencies; (vi) assists Chief Learning Officers in selecting the most appropriate employee for a given role; (vii) assists a Project Manager in building teamwork. A prototype tool implementing MASEL using JADE (Java Agent DEvelopment Framework) was developed. The reasoning capability of MASEL agents involved in the learning paths building process and in the team building process is implemented using DLV, a disjunctive logic programming system.

Decomposition of planning problems

Abstract: The ability to decompose a problem into manageable sub-components is a necessity in complex problem-solving tasks. In planning, the application of a divide-and-conquer methodology is known as planning decomposition. This technique consists of the following: decomposing a problem into smaller components (subproblems), solving these subproblems individually, and then combining the obtained solutions. The success of this technique is subject to the interactions that may appear between actions from solutions for different subproblems.In this paper, we present a novel technique, STeLla, to overcome the inherent difficulties in planning decomposition. This technique partitions a planning problem in such a way that its subproblems can then be solved separately (either sequentially or concurrently) and their solutions can be easily combined. The key issue is that interactions among goals are used to come up with the problem decomposition rather than solving them once the problem is decomposed. This approach proves to be very beneficial with respect to other decomposition methods and state-of-the-art planners.

On-line monitoring and diagnosis of a team of service robots: A model-based approach

Abstract: The paper presents an approach for the on-line monitoring and diagnosis of multi-robot systems where services are provided by a team of robots and the environment is only partially observable via a net of fixed sensors. This kind of systems exhibits complex dynamics where weakly predictable interactions among robots may occur. To face this problem, a model-based approach is adopted: in particular, the paper discusses how to build a system model by aggregating a convenient set of basic system components, which are modeled via communicating automata. Since the dynamics of a multi-robot system depend on the actions performed by the robots (and actions change over time), the global system model is partitioned into a number of submodels, each one describing the dynamics of a single action.The paper introduces the architecture of the Supervisor which has to track the actions progress and to infer an explanation when an action is completed with delay or fails. The Supervisor includes two main modules: the On-line Monitoring Module (OMM) tracks the status of the system by exploiting the (partial) observations provided by sensors and robots. When the monitor detects failures in the actions execution, the Diagnostic Interpretation Module (DIM) is triggered for explaining the failure in terms of faults in the robots and/or troublesome interactions among them.The RoboCare domain has been selected as a test bed of the approach. The paper discusses experimental results collected in such a domain with particular focus on the competence and the efficiency of both the OMM and the DIM.

Agents' roles in B2C e-commerce

Abstract: The rapid evolution of the Internet from a general information space to an electronic market space provides the users with the possibility to navigate among thousands of Web sites for comparing products and merchants, for making their purchases or for obtaining some desired services. However, from the user's viewpoint, such a navigation often requires an high cost in terms of time to spend on the Web to perform a satisfying comparison of the various alternatives. From the supplier's viewpoint, it is needed to propose products in a suitable way to customers, taking into account the typology of each customer, her/his preferences, habits, etc.In this context, generally denoted as Business-to-Customer (B2C) e-commerce, the use of software agents as mediators in e-commerce activities seems to be particularly promising.In this paper, we describe and analyze the various roles agents have assumed in B2C e-commerce applications. Furthermore, we propose a Consumer Buying Behaviour model, called E2-CBB, that considers new emergent issues, as the capability to solve semantic heterogeneity, and the adaptive presentations of Web stores. By using such a model, we classify and compare a number of agent-based approaches for managing B2C e-commerce, proposed in the literature in the last ten years.

Template programs for Disjunctive Logic Programming: An operational semantics

Abstract: Disjunctive Logic Programming is nowadays a mature formalism which has been successfully applied to a variety of practical problems, such as information integration, knowledge representation, planning, diagnosis, optimization and configuration. Although current DLP systems have been extended in many directions, they still miss features which may be helpful towards industrial applications, like the capability of quickly introducing new predefined constructs or of dealing with modules. Indeed, in spite of the fact that a wide literature about modular logic programming is known, code reusability has never been considered as a critical point in Disjunctive Logic Programming. In this work we extend Disjunctive Logic Programming, under stable model semantics, with the notion of "template" predicates. A template predicate may be instantiated to an ordinary predicate by means of template atoms, thus allowing to define reusable modules, to define new constructs and aggregates without any syntactic limitation.

A conceptual clustering approach for user profiling in personal information agents

Abstract: Information agents have emerged in the last decade as an alternative to assist users to cope with the increasing volume of information available on the Web. In order to provide personalized assistance, these agents rely on having some knowledge about users contained into user profiles, i.e., models of users preferences and interests gathered by observation of user behavior. User profiles have to summarize categories corresponding not only to diverse user information interests but also to different levels of abstraction in order to allow agents to decide on the relevance of new pieces of information. In accomplishing this goal, the discovery of interest categories using document clustering offers the advantage that an a priori knowledge of user interests is not needed, therefore the process of acquiring profiles is completely unsupervised. However, most document clustering algorithms are not applicable to the problem of incrementally acquiring and modeling interests because of either the kind of solutions they provide, which do not resemble user interests, or the way they build such solutions, which is generally not incremental. In this paper we describe and evaluate a document clustering algorithm, named WebDCC (Web Document Conceptual Clustering), designed to support learning of user interests by personal information agents. WebDCC algorithm carries out incremental, unsupervised concept learning over Web documents with the goal of building and maintaining both accurate and comprehensible user profiles. Empirical evaluation of using this algorithm for user profiling and its advantages with respect to other clustering algorithms are presented.

The CADE-20 automated theorem proving competition

Abstract: The CADE ATP System Competition (CASC) is an annual evaluation of fully automatic, first order Automated Theorem Proving systems CASC-20 was the tenth competition in the CASC series Seventeen ATP systems and system variants competed in the various competition and demonstration divisions An outline of the competition design, and a commentated summary of the results, are presented

Tools for modeling and solving search problems

Abstract: In this paper, we describe a language PSpb to model search problems that are specified in terms of boolean combinations of pseudo-boolean constraints. We then describe software tools that allow one to use SAT and SAT(PB) solvers to compute solutions to instances of search problems represented in the language PSpb.

Automated reasoning in quantified modal and temporal logics

Abstract: This paper is a summary of the author's PhD thesis, concerned with automated reasoning in quantified modal and temporal logics. The relevant contributions are: (i) a sound and complete set of sequent calculi for quantified modal logics is devised; (ii) the approach is extended to the quantified temporal logic of linear, discrete time and a framework for doing automated reasoning via Proof Planning in it is developed; (iii) a set of promising experimental results is shown, obtained by applying the framework to the problem of Feature Interactions in telecommunication systems.

Rewriting queries using views with negation

Abstract: Data integration and query reformulation are classical examples of problems that require techniques developed in both AI and database systems fields In this work we address the problem of rewriting queries using views which has many applications In particular, we consider queries and views that are conjunctive queries with safe negation (CQNs) We prove that given a CQN query and a set of CQN views, finding equivalent rewritings is decidable in both cases where the rewriting is in the language of CQNs or unions of CQNs We prove decidability by giving upper bounds on the number of subgoals of the rewritings We limit the search space of potential equivalent CQN rewritings and Maximally Contained CQN Rewritings (MCRs) to the language of unions of CQs in the case where the query is CQ Finally, we give a sound and complete algorithm for finding equivalent rewritings in the language of unions of CQNs and we prove that if we consider the rewritings without negated subgoals, then they compute only certain answers under the OWA Of independent interest is a simple test for checking containment between two unions of CQNs

Evolutionary concept learning in first order logic: an overview

Abstract: This paper presents an overview of evolutionary approaches to Inductive Logic Programming (ILP). After a short description of the two popular ILP systems FOIL and Progol, we focus on methods based on evolutionary algorithms (EAs). Six systems are described and compared by means of the following aspects: search strategy, representation, hypothesis evaluation, search operators and biases adopted for limiting the hypothesis space. We discuss possible advantages and drawbacks related to the specific features of the systems along these aspects. Issues concerning the relative performance and efficiency of the systems are addressed.

Fuzzy constraint satisfaction approach for landmark recognition in mobile robotics

Abstract: This work deals with landmark recognition in mobile robotics, using a new model based on Constraint Satisfaction Problems (CSP): the Multivariable Fuzzy Temporal Profile model (MFTP). A representation supported by CSPs makes it possible to capture a morphological description of the patterns that landmarks give rise to on sensor readings. Its representation, based on Fuzzy Set Theory, allows the imprecision and uncertainty that are characteristic of the problem to be handled. The work places special emphasis on those aspects that are resolved by means of this approach: the ability to model semantically rich landmarks, the simplicity of its description, and the high computational efficiency of the proposed detection algorithms. Finally, a validation of the model in the detection of various landmarks over ultrasound (US) sensors is presented. In spite of these sensors being highly noisy and imprecise, the MFTP model successfully detects 95p of the landmarks on the reference wall.

An inductive logic programming approach to statistical relational learning: Thesis

Abstract: Statistical relational learning (SRL) addresses one of the central open questions of AI: the combination of relational or first-order logic with principled probabilistic and statistical approaches to inference and learning This thesis approaches SRL from an inductive logic programming (ILP) perspective and starts with developing a general framework for SRL: probabilistic ILP Based on this foundation, the thesis shows how to incorporate the logical concepts of objects and relations among these objects into Bayesian networks As time and actions are not just other relations, it afterwards develops approaches to probabilistic ILP over time and for making complex decision in relational domains Finally, it is shown that SRL approaches naturally yield kernels for structured data The resulting approaches are illustrated using examples from genetics, bioinformatics, and planning domains

A guided learning strategy for vision based navigation of 4-legged robots

Abstract: This paper presents a new approach to building low level navigation behaviors for 4-legged robots through vision based demonstration learning. The main novelty of the approach is that rather than observing other entities and adapting their kinematics to the robot constraints, a supervisor controls the robot to achieve the desired behavior through a proper interface. The guided actions and the relevant input parameters are related via Case Base Reasoning, so that the robot can retrieve them later to work in an unsupervised way. This intuitive acquisition of reactive behaviors allows bottom-up construction of more complex emergent behaviors and avoids low level kinematics analysis and possible associated errors. The system has been successfully tested using a Sony Aibo robot. Experiments have proven that the robot is capable of adopting a variety of reliable behaviors depending on its relative position in relation to a ball through different trainings. Also, being reactive, the system is resistant against punctual errors and occlusions.

Fault tolerant knowledge level inter-agent communication in open Multi-Agent Systems: Thesis

Abstract: In the thesis we propose FT-ACL, an advanced Agent Communication Language which deals with crash failures of agents running in an open and dynamic MAS. FT-ACL provides high-level fault-tolerant communication primitives and support for an anonymous interaction protocol designed for open MAS. Moreover, the ACL satisfies a set of well defined knowledge-level programming requirements. In the thesis we provide a formal semantics for FT-ACL and a formal specification of the underlying agent architecture.

Learning user interests for user profiling in personal information agents: Thesis

Abstract: This work presents a novel user profiling technique, named WebProfiler, developed to support incremental learning and adaptation of user profiles for agents assisting users with information-related tasks. This technique aims at acquiring comprehensible user profiles that accurately capture user interests starting from the observation of user behavior on the Web.

A conceptual clustering approach for user profiling in personal information agents

Abstract: Information agents have emerged in the last decade as an alternative to assist users to cope with the increasing volume of information available on the Web. In order to provide personalized assista...

Evolutionary concept learning in first order logic

Abstract: This paper presents an overview of evolutionary approaches to Inductive Logic Programming (ILP). After a short description of the two popular ILP systems FOIL and Progol, we focus on methods based ...

Domain-independent temporal planning in a planning-graph-based approach

Abstract: Many planning domains have to deal with temporal features that can be expressed using durations that are associated to actions. This paper presents a temporal planning approach that combines the principles of Graphplan and TGP, and uses the information calculated in the planning graph to deal with a model of actions that include local conditions and effects. In this approach, we propose two strategies for search. The first one is based on the Graphplan backward search. The second one is a new strategy for search based on a least-commitment and heuristic search, which attempts to overcome the main limitations of a chronological backtracking search when dealing with large temporal problems. This search has proved to be beneficial in the scalability of the planner and the experiments show that a planner using this new search is competitive with other state-of-the-art planners w.r.t. the plan quality.

Integrating heterogeneous adaptation techniques to build a flexible and usable mobile tourist guide

Abstract: Intelligent adaptation is a key issue for the design of flexible support systems for mobile users. In this paper we present UbiquiTO, a tourist guide which integrates different forms of adaptation:...

A Compact Representation for Least Common Subsumers in the description logic ALE

Abstract: This paper introduces a compact representation which helps to avoid the exponential blow-up in space of the Least Common Subsumer (lcs) of two ALE-concept descriptions. Based on the compact representation we define a space of specific graphs which represents all ALE-concept descriptions including the lcs. Next, we propose an algorithm exponential in time and polynomial in space for deciding subsumption between concept descriptions represented by graphs in this space. These results provide better understanding of the double exponential blow-up of the approximation of ALC-concept descriptions by ALE-concept descriptions: double exponential size of the approximation in the ordinary representation is unavoidable in the worst case.

Propositional satisfiability: techniques, algorithms and applications

Abstract: Recent years have seen remarkable progress in propositional satisfiability (SAT) Despite the worst-case exponential run time of all known algorithms, SAT solvers can currently be used to solve hard benchmark problemsThis PhD dissertation contributes to a better understanding of the techniques, the algorithms and the applications of propositional satisfiability First, we introduce efficient lazy data structures that, though unable to determine exactly the dynamic size of a clause, are quite accurate at predicting the number of unassigned literals in a clause In addition, we suggest the use of probing-based preprocessing techniques for manipulating propositional formulas Furthermore, unrestricted backtracking is proposed as an algorithm that combines the advantages of local search and backtrack search Finally, we relate hardness with hidden structure in unsatisfiable random 3-SAT formulas, where hardness is measured by the search effort and hidden structure is measured by unsatisfiable cores and strong backdoors

Fuzzy constraint satisfaction approach for landmark recognition in mobile robotics

Abstract: This work deals with landmark recognition in mobile robotics, using a new model based on Constraint Satisfaction Problems (CSP): the Multivariable Fuzzy Temporal Profile model (MFTP). A representat...