Domenico Corapi

Imperial College London, Londinium, England, United Kingdom

Are you Domenico Corapi?

Claim your profile

Publications (14)1.05 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: Recent work has shown how a meta-level approach to inductive logic programming, which uses a semantic-preserving transformation of a learning task into an abductive reasoning problem, can address a large class of multi-predicate, nonmonotonic learning in a sound and complete manner. An Answer Set Programming (ASP) implementation, called ASPAL, has been proposed that uses ASP fixed point computation to solve a learning task, thus delegating the search to the ASP solver. Although this meta-level approach has been shown to be very general and flexible, the scalability of its ASP implementation is constrained by the grounding of the meta-theory. In this paper we build upon these results and propose a new meta-level learning approach that overcomes the scalability problem of ASPAL by breaking the learning process up into small manageable steps and using theory revision over the meta-level representation of the hypothesis space to improve the hypothesis computed at each step. We empirically evaluate the computational gain with respect to ASPAL using two different answer set solvers.
    No preview · Chapter · Jan 2014
  • [Show abstract] [Hide abstract]
    ABSTRACT: Environment domain models are a key part of the information used by adaptive systems to determine their behaviour. These models can be incomplete or inaccurate. In addition, since adaptive systems generally operate in environments which are subject to change, these models are often also out of date. To update and correct these models, the system should observe how the environment responds to its actions, and compare these responses to those predicted by the model. In this paper, we use a probabilistic rule learning approach, NoMPRoL, to update models using feedback from the running system in the form of execution traces. NoMPRoL is a technique for nonmonotonic probabilistic rule learning based on a transformation of an inductive logic programming task into an equivalent abductive one. In essence, it exploits consistent observations by finding general rules which explain observations in terms of the conditions under which they occur. The updated models are then used to generate new behaviour with a greater chance of success in the actual environment encountered.
    No preview · Conference Paper · May 2013
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Normative frameworks provide a means to address the governance of open systems, by offering a mechanism to express responsibilities and permissions of the individual participants with respect to the entire system without compromising their autonomy. Careful design is crucial if it is to meet its requirements. Tools that support the design process can be of great benefit. In this paper, we describe a method for choosing the appropriate change in the normative specification, using impact analysis of the critical consequences being preserved or rejected by the change.
    Full-text · Conference Paper · Jun 2012
  • Source
    Domenico Corapi · Alessandra Russo · Emil Lupu
    [Show abstract] [Hide abstract]
    ABSTRACT: In this paper we discuss the design of an Inductive Logic Programming (ILP) system in Answer Set Programming (ASP) and more in general the problem of integrating the two. We show how to formalise the learning problem as an ASP program and provide details on how the optimisation features of modern solvers can be adapted to derive preferred hypotheses.
    Full-text · Conference Paper · Jul 2011
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: In this paper we propose a use-case-driven iterative design methodology for normative frameworks, also called virtual institutions, which are used to govern open systems. Our computational model represents the normative framework as a logic program under answer set semantics (ASP). By means of an inductive logic programming approach, implemented using ASP, it is possible to synthesise new rules and revise the existing ones. The learning mechanism is guided by the designer who describes the desired properties of the framework through use cases, comprising (i) event traces that capture possible scenarios, and (ii) a state that describes the desired outcome. The learning process then proposes additional rules, or changes to current rules, to satisfy the constraints expressed in the use cases. Thus, the contribution of this paper is a process for the elaboration and revision of a normative framework by means of a semi-automatic and iterative process driven from specifications of (un)desirable behaviour. The process integrates a novel and general methodology for theory revision based on ASP.
    Full-text · Article · Jul 2011 · Theory and Practice of Logic Programming
  • Domenico Corapi · Daniel Sykes · Katsumi Inoue · Alessandra Russo
    [Show abstract] [Hide abstract]
    ABSTRACT: We propose here a novel approach to rule learning in probabilistic nonmonotonic domains in the context of answer set programming. We used the approach to update the knowledge base of an agent based on observations. To handle the probabilistic nature of our observation data, we employ parameter estimation to find the probabilities associated with each of these atoms and consequently with rules. The outcome is the set of rules which have the greatest probability of entailing the observations. This ultimately improves tolerance of noisy data compared to traditional inductive logic programming techniques. We illustrate the benefits of the approach by applying it to a planning problem in which the involved agent requires both nonmonotonicity and tolerance of noisy input.
    No preview · Conference Paper · Jul 2011
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Discovering the Business Process (BP) model underpinning existing practices through analysis of event logs, allows users to understand, analyse and modify the process. But, to be useful, the BP model must be kept in line with practice throughout its lifetime, as changes occur to the business objectives, technologies and quality programs. Current techniques require users to manually revise the BP to account for discrepancies between the practice and the model, which is a laborious, costly and error prone task. We propose an automated approach for resolving such discrepancies by minimally revising a BP model to bring it in line with the activities corresponding to its executions, based on a non-monotonic inductive learning system. We discuss our implementation of this approach and demonstrate its application to a case-study. We further contrast our approach with existing BP discovery techniques to show that BP revision offers significant advantages over BP discovery in practical use.
    Full-text · Conference Paper · Sep 2010
  • Source
    Ryan Wishart · Domenico Corapi · Srdjan Marinovic · Morris Sloman
    [Show abstract] [Hide abstract]
    ABSTRACT: Recent years have seen a significant increase in the popularity of social networking services. These online services enable users to construct groups of contacts, referred to as friends, with which they can share digital content and communicate. This sharing is actively encouraged by the social networking services, with users' privacy often seen as a secondary concern. In this paper we first propose a privacy-aware social networking service and then introduce a collaborative approach to authoring privacy policies for the service. In addressing user privacy, our approach takes into account the needs of all parties affected by the disclosure of information and digital content.
    Full-text · Conference Paper · Aug 2010
  • Source
    Ryan Wishart · Domenico Corapi · Anil Madhavapeddy · Morris Sloman
    [Show abstract] [Hide abstract]
    ABSTRACT: The online presence projected by a person is comprised of all the information about them available on the Internet. In online communities and social networking services, it is often possible for third-parties to modify this content by, for example, commenting on existing content or uploading new content. This has the potential to negatively impact the privacy of a presence owner (the person referred to by the on-line content) by disclosing information about them without consent. In this paper we propose a Privacy Butler, an automated service that can monitor a person's online presence and attempt to make corrections based on policies specified by the owner of the online presence.
    Full-text · Conference Paper · May 2010
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: In the physical world, the rules governing behaviour are debugged by observing an outcome that was not intended and the addition of new constraints to prevent the attainment of that outcome. We propose a similar approach to support the incremental development of normative frameworks (also called institutions) and demonstrate how this works through the validation and synthesis of normative rules using model generation and inductive learning. This is achieved by the designer providing a set of use cases, comprising collections of event traces that describe how the system is used along with the desired outcome with respect to the normative framework. The model generator encodes the description of the current behaviour of the system. The current specification and the traces for which current behaviour and expected behaviour do not match are given to the learning framework to propose new rules that revise the existing norm set in order to inhibit the unwanted behaviour. The elaboration of a normative system can then be viewed as a semi-automatic, iterative process for the detection of incompleteness or incorrectness of the existing normative rules, with respect to desired properties, and the construction of potential additional rules for the normative system.
    Full-text · Conference Paper · Jan 2010
  • Source
    Domenico Corapi · Alessandra Russo · Emil Lupu
    [Show abstract] [Hide abstract]
    ABSTRACT: We present a novel approach to non-monotonic ILP and its implementation called TAL (top-directed abductive learning). TAL overcomes some of the completeness problems of ILP systems based on inverse entailment and is the first top-down ILP system that allows background theories and hypotheses to be normal logic programs. The approach relies on mapping an ILP problem into an equivalent ALP one. This enables the use of established ALP proof procedures and the specification of richer language bias with integrity constraints. The mapping provides a principled search space for an ILP problem, over which an abductive search is used to compute inductive solutions.
    Full-text · Conference Paper · Jan 2010
  • Source
    Domenico Corapi · Oliver Ray · Alessandra Russo · Arosha Bandara · Emil Lupu
    [Show abstract] [Hide abstract]
    ABSTRACT: Pervasive computing requires infrastructures that adapt to changes in user behaviour while minimising user interactions. Policy-basedapproaches have been proposed as a means of providing adaptability but, at present, require policy goals and rules to be explicitly defined by users. This paper presents a novel, logic-based approach for automatically learning and updating models of users from their observed behaviour. We show how this task can be accomplished using a nonmonotonic learning system, and we illustrate how the approach can be exploited within a pervasive computing framework.
    Full-text · Article · Apr 2009 · IFIP International Federation for Information Processing
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: With the emergence of ubiquitous computing, innovations in mobile phones are increasingly changing the way users lead their lives. To make mobile devices adaptive and able to autonomously respond to changes in user behaviours, machine learning techniques can be deployed to learn behaviour from empirical data. Learning outcomes should be rule-based enforcement policies that can manage pervasively the devices, and at the same time facilitate user validation when and if required. In this paper we demonstrate the feasibility of non-monotonic ILP in the automated task of extraction of user behaviour rules through data acquisition in the domain of mobile phones. This is a challenging task as real mobile datasets are highly noisy and unevenly distributed. We present two applications, one based on an existing dataset collected as part of the Reality Mining group, and the other generated by a mobile phone application, called ULearn, that we have developed to facilitate a realistic evaluation of the accuracy of the learning outcome.
    Full-text · Article ·
  • Source
    Domenico Corapi · Alessandra Russo
    [Show abstract] [Hide abstract]
    ABSTRACT: We provide here a brief introduction and proof of soundness and completeness of the ILP system ASPAL. This document is in support of our ICLP 2011 submission, for the reviewers' benefits. Inductive Logic Programming (ILP) [4] is a machine learning technique con-cerned with the induction of logic theories, called target theories, that generalise (positive and negative) examples with respect to a prior background knowledge. For example, from the observations f ly(a), f ly(b), ¬f ly(c) and a background knowledge containing the two facts bird(a) and bird(b), we can generalise the concept f ly(X) ← bird(X). In non-trivial problems it is crucial to define the space of possible solutions accurately. Target theories are within a space defined by a language bias, which can be expressed using mode declarations [4]. We refer to [3] for notations and preliminary definitions on logic program-ming. For a given theory T we denote with B T the Herbrand base of T . Definition 1. A mode declaration is either a head declaration, written modeh(s), or a body declaration, written modeb(s), where s is a schema. A schema is a ground literal containing special terms called placemarkers. A placemarker is ei-ther '+type', '−type' or '#type' where type denotes the type of the placemarker and the three symbols '+', '−' and '#' indicate that the placemarker is an input, an output, a constant placemarker, respectively.
    Full-text · Article ·