This paper presents GRT, a domain-independent heuristic planning system for STRIPS worlds. GRT so... more This paper presents GRT, a domain-independent heuristic planning system for STRIPS worlds. GRT solves problems in two phases. In the pre-processing phase, it estimates the distance between each fact and the goals of the problem, in a backward direction. Then, in the search phase, these estimates are used in order to further estimate the distance between each intermediate state and the goals, guiding so the search process in a forward direction and on a best-first basis. The paper presents the benefits from the adoption of opposite directions between the preprocessing and the search phases, discusses some difficulties that arise in the pre-processing phase and introduces techniques to cope with them. Moreover, it presents several methods of improving the efficiency of the heuristic, by enriching the representation and by reducing the size of the problem. Finally, a method of overcoming local optimal states, based on domain axioms, is proposed. According to it, difficult problems are ...
In many practical applications of supervised learning the task involves the prediction of multipl... more In many practical applications of supervised learning the task involves the prediction of multiple target variables from a common set of input variables. When the prediction targets are binary the task is called multi-label classification, while when the targets are continuous the task is called multi-target regression. In both tasks, target variables often exhibit statistical dependencies and exploiting them in order to improve predictive accuracy is a core challenge. A family of multi-label classification methods address this challenge by building a separate model for each target on an expanded input space where other targets are treated as additional input variables. Despite the success of these methods in the multi-label classification domain, their applicability and effectiveness in multi-target regression has not been studied until now. In this paper, we introduce two new methods for multi-target regression, called stacked single-target and ensemble of regressor chains, by adapting two popular multilabel classification methods of this family. Furthermore, we highlight an inherent problem of these methods-a discrepancy of the values of the additional input variables between training and prediction-and develop extensions that use out-of-sample estimates of the target variables during training in order to tackle this problem. The results of an extensive experimental evaluation carried out on a large and diverse collection of datasets show that, when the discrepancy is appropriately mitigated, the proposed methods attain consistent improve-Editor: Johannes Fürnkranz.
The management of large data networks, like a national WAN, is without any doubt a complex task. ... more The management of large data networks, like a national WAN, is without any doubt a complex task. Taking into account the constantly increasing size and complexity of today's TCP/IP based networks, it becomes obvious that there is a demanding need for better than simple monitoring management tools. Expert system technology seems to be a very promising approach for the development of such tools. This paper describes the system architecture of ExperNet, a distributed expert system for the management of the National Computer Network of Ukraine, and the implementation of the tools used for its development. ExperNet is a multiagent system built in DEVICE, an active OODB enhanced with high level rules, that uses CS-Prolog II to implement the communication facilities required. The system employs HNMS+ and Big-Brother, two modi ed versions of existing network management tools, in order to obtain a complete view of the monitored network.
This paper proposes a framework for educational software evaluation based on the Multiple Criteri... more This paper proposes a framework for educational software evaluation based on the Multiple Criteria Decision Aid methodology. Evaluating educational software products is a twofold process: both the educational and the technical aspect of the evaluated products have to be considered. As far as the product educational effectiveness is concerned, we propose a set of attributes covering both the general educational features and the content of the product. From the technical point of view, a software attribute set based on the ISO/IEC 9126 standard has been chosen together with the accompanying measurement guidelines. Finally, an evaluation example involving three commercial educational software packages for mechanics is presented.
This chapter is concerned with the enhancement of planning systems using techniques from Machine ... more This chapter is concerned with the enhancement of planning systems using techniques from Machine Learning in order to automatically configure their planning parameters according to the morphology of the problem in hand. It presents two different adaptive systems that set the planning parameters of a highly adjustable planner based on measurable characteristics of the problem instance. The planners have acquired their knowledge from a large data set produced by results from experiments on many problems from various domains. The first planner is a rulebased system that employs propositional rule learning to induce knowledge that suggests effective configuration of planning parameters based on the problem's characteristics. The second planner employs instance-based learning in order to find problems with similar structure and adopt the planner configuration that has proved in the past to be effective on these problems. The validity of the two adaptive systems is assessed through experimental results that demonstrate the boost in performance in problems of both known and unknown domains. Comparative experimental results for the two planning systems are presented along with a discussion of their advantages and disadvantages.
E-mail has met tremendous popularity over the past few years. People are sending and receiving ma... more E-mail has met tremendous popularity over the past few years. People are sending and receiving many messages per day, communicating with partners and friends or exchanging files and information. Unfortunately, the phenomenon of e-mail overload has grown over the past years, becoming a personal headache for users and a financial issue for companies. In this chapter, we will discuss how disciplines like machine learning and data mining can contribute to the solution of the problem by constructing intelligent techniques that automate e-mail managing tasks and what advantages they hold over other conventional solutions. We will also discuss the particularity of e-mail data and what special treatment they require. Some interesting e-mail mining applications like mail categorization, summarization, automatic answering, and spam filtering will also be presented.
In this paper we present R-DEVICE, a deductive rule language for reasoning about RDF metadata. R-... more In this paper we present R-DEVICE, a deductive rule language for reasoning about RDF metadata. R-DEVICE includes features such as normal and generalized path expressions, stratified negation, aggregate, grouping, and sorting, functions. The rule language supports a second-order syntax, where variables can range over classes and properties. Users can define views which are materialized and, optionally, incrementally maintained by translating deductive rules into CLIPS production rules. Users can choose between an OPS5/CLIPS-like or a RuleML-like syntax. R-DEVICE is based on a OO RDF data model, different than the established graph model, which maps resources to objects and encapsulates properties inside resource objects, as traditional OO attributes. In this way, less joins are required to access the properties of a single resource resulting in better inferencing/querying performance. The descriptive semantics of RDF may call for dynamic re-definitions of resource classes and objects, which are handled by R-DEVICE effectively.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2005
Defeasible reasoning is a rule-based approach for efficient reasoning with incomplete and inconsi... more Defeasible reasoning is a rule-based approach for efficient reasoning with incomplete and inconsistent information. Such reasoning is useful for many applications in the Semantic Web, such as policies and business rules, agent brokering and negotiation, ontology and knowledge merging, etc. However, the syntax of defeasible logic may appear too complex for many users. In this paper we present a graphical authoring tool for defeasible logic rules that acts as a shell for the DR-DEVICE defeasible reasoning system over RDF metadata. The tool helps users to develop a rule base using the OO-RuleML syntax of DR-DEVICE rules, by constraining the allowed vocabulary through analysis of the input RDF namespaces, so that the user does not have to type-in class and property names. Rule visualization follows the tree model of RuleML. The DR-DEVICE reasoning system is implemented on top of the CLIPS production rule system and builds upon an earlier deductive rule system over RDF metadata that also supports derived attribute and aggregate attribute rules.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2010
Public institutes often face the challenge of managing vast volumes of administrative documents, ... more Public institutes often face the challenge of managing vast volumes of administrative documents, a need that is often met via Content Management Systems (CMSs). CMSs offer various advantages, like separation of data structure from presentation and variety in user roles, but also present certain disadvantages, like inefficient keyword-based search facilities. The new generation of content management solutions imports the notion of semantics and is based on Semantic Web technologies, such as metadata and ontologies. The benefits include semantic interoperability, competitive advantages and dramatic cost reduction. In this paper a leading Enterprise CMS is extended with semantic capabilities for automatically importing and exporting ontologies. This functionality enables reuse of repository content, semantically-enabled search and interoperability with third-party applications. The extended system is deployed in semantically managing the large volumes of documents for a state university.
Proceedings of the 3rd International Conference on Web Intelligence, Mining and Semantics, 2013
Although the Semantic Web and Web Service technologies have already formed a synergy towards Sema... more Although the Semantic Web and Web Service technologies have already formed a synergy towards Semantic Web Services, their use remains limited. Potential adopters are usually discouraged by the number of different methodologies and the lack of tools, which both force them to acquire expert knowledge and commit to exhausting manual labor. This work proposes a novel functional and userfriendly graphical tool, named Iridescent, intended for both expert and non-expert users, to create and edit Semantic Web Service descriptions, following the SAWSDL recommendation. The tool"s aim is twofold: to enable users manually create descriptions in a visual manner, providing a complete alternative to coding, and to semiautomate the process by matching elements and concepts and suggesting annotations. A state-of-the-art survey has been carried out to reveal critical points and requirements. The tool"s functionality is presented along with usage scenarios that demonstrate how the tool and SAWSDL enable Intelligence in an Ambient Intelligence environment. Finally, Iridescent was methodically tested for its usability and evaluated by a range of both expert and non-expert users.
Solving software evaluation problems is a particularly difficult software engineering process and... more Solving software evaluation problems is a particularly difficult software engineering process and many different-often contradictory-criteria must be considered in order to reach a decision. This paper presents ESSE, a prototype expert system for software evaluation that embodies various aspects of the Multiple-Criteria Decision Aid (MCDA) methodology. Its main features are the flexibility in problem modeling and the built-in knowledge about software problem solving and software attribute assessment. Evaluation problems are modeled around top-level software attributes, such as quality and cost. Expert assistants guide the evaluator in feeding values to the decision model. ESSE covers all important dimensions of software evaluation through the integration of different technologies.
International Journal on Artificial Intelligence Tools, 2007
Rare events analysis is an area that includes methods for the detection and prediction of events,... more Rare events analysis is an area that includes methods for the detection and prediction of events, e.g. a network intrusion or an engine failure, that occur infrequently and have some impact to the system. There are various methods from the areas of statistics and data mining for that purpose. In this article we propose PREVENT, an algorithm which uses inter-transactional patterns for the prediction of rare events in transaction databases. PREVENT is a general purpose inter-transaction association rules mining algorithm that optimally fits the demands of rare event prediction. It requires only 1 scan on the original database and 2 over the transformed, which is considerably smaller and it is complete as it does not miss any patterns. We provide the mathematical formulation of the problem and experimental results that show PREVENT's efficiency in terms of run time and effectiveness in terms of sensitivity and specificity.
Solving software evaluation problems is a particularly difficult software engineering process and... more Solving software evaluation problems is a particularly difficult software engineering process and many contradictory criteria must be considered to reach a decision. Nowadays, the way that decision support techniques are applied suffers from a number of severe problems, such as naive interpretation of sophisticated methods and generation of counter-intuitive, and therefore most probably erroneous, results. In this paper we identify some common flaws in decision support for software evaluations. Subsequently, we discuss an integrated solution through which significant improvement may be achieved, based on the Multiple Criteria Decision Aid methodology and the exploitation of packaged software evaluation expertise in the form of an intelligent system. Both common mistakes and the way they are overcomed are explained through a real world example.
IEEE Transactions on Knowledge and Data Engineering, 2003
This paper describes the integration of a multidatabase system and a knowledge-base system to sup... more This paper describes the integration of a multidatabase system and a knowledge-base system to support the data-integration component of a Data Warehouse. The multidatabase system integrates various component databases with a common query language, however it does not provide capability for schema integration and other utilities necessary for Data Warehousing. The knowledge base system offers in addition a declarative logic language with second-order syntax but first-order semantics for integrating the schemes of the data sources into the warehouse and for defining complex, recursively defined materialized views. Furthermore, deductive rules are also used for cleaning, checking the integrity and summarizing the data imported into the Data Warehouse. The Knowledge Base System features an efficient incremental view maintenance mechanism that is used for refreshing the Data Warehouse, without querying the data sources.
This paper describes ExperNet, an intelligent multi-agent system that was developed under an EU f... more This paper describes ExperNet, an intelligent multi-agent system that was developed under an EU funded project to assist in the management of a large-scale data network. ExperNet assists network operators at various nodes of a WAN to detect and diagnose hardware failures and network traffic problems and suggests the most feasible solution, through a web-based interface. ExperNet is composed by intelligent agents, capable of both local problem solving and social interaction among them for coordinating problem diagnosis and repair. The current network state is captured and maintained by conventional network management and monitoring software components, which have been smoothly integrated into the system through sophisticated information exchange interfaces. For the implementation of the agents, a distributed Prolog system enhanced with networking facilities was developed. The agents' knowledge base is developed in an extensible and reactive knowledge base system capable of handling multiple types of knowledge representation. ExperNet has been developed, installed and tested successfully in an experimental network zone of Ukraine.
FUNAGES is an expert system that deals with the interpretation of fundus fluorescein angiography.... more FUNAGES is an expert system that deals with the interpretation of fundus fluorescein angiography. Fluorescein angiography is an extremely valuable clinical test that provides information about the circulatory system of the ocular fundus (the back of the eye) not attainable by routine examination. The different appearance of fluorescein, in place and time and the classification of the fundus diseases render angiography a dynamic, cinematographic and deductive diagnostic method. Therefore, ability to interpret fundus fluorescein angiograms allows an ophthalmologist specializing in ocular fundus diseases to follow a systematic, orderly and logical line of reasoning that leads to a proper diagnosis. FUNAGES was developed to simulate such logical reasoning, in order to train inexperienced ophthalmologists in the interpretation of angiograms. The system achieved its purpose via a graphical user interface and a thorough knowledge base.
This study examines the presence of either linear or nonlinear relationships between a number of ... more This study examines the presence of either linear or nonlinear relationships between a number of popular seawater quality indicators such as water temperature, pH, amount of dissolved oxygen and turbidity. The data are obtained from a set of sensors in an underwater measurement station. The neural networks with active neurons are applied to the prediction of each one of the above four indicators and their performance is compared against a benchmark prediction method known as the random walk model. The random walk model is the simpler prediction method, which accepts as the best prediction for a variable its current value. The neural network with active neurons is a black box method, which contrary to neural networks with passive neurons does not require a long set of training data. The results show that for daily predictions the neural network with active neurons is able to beat the random walk model with regard to directional accuracy, namely the direction (upward or downwards) of the modelling object in the next day.
In this paper we apply Regression via Classification (RvC) to the problem of estimating the numbe... more In this paper we apply Regression via Classification (RvC) to the problem of estimating the number of software defects. This approach apart from a certain number of faults, it also outputs an associated interval of values, within which this estimate lies with a certain confidence. RvC also allows the production of comprehensible models of software defects exploiting symbolic learning algorithms. To evaluate this approach we perform an extensive comparative experimental study of the effectiveness of several machine learning algorithms in two software data sets. RvC manages to get better regression error than the standard regression approaches on both datasets.
This work presents a Web Service Middleware infrastructure for Ambient Intelligence environments,... more This work presents a Web Service Middleware infrastructure for Ambient Intelligence environments, named aWESoME. aWESoME is a vital part of the Smart IHU project, a large-scale Smart University deployment. The purpose of the proposed middleware within the project is twofold: for one, to ensure universal, homogeneous access to the system's functions and secondly, to fulfill functional and non-functional requirements of the system. Namely, the infrastructure itself should consume significantly low power (as it is meant for energy savings in addition to automations), without compromising reliability and fast response time. The infrastructure should enable fast and direct discovery, invocation and execution of services. Finally, on hardware level, the wireless sensor and actuator network should be optimally configured for speed and reliability as well. The proposed solution employs widely used web open standards for description and discovery to expose hardware and software functions and ensure interoperability, even outside the borders of this university deployment. It proposes a straightforward method to integrate low-cost and resource-constrained heterogeneous devices found in the market and a largescale placement of servers and wireless sensor networks. Different server hardware installations have been evaluated to find the optimum trade-off between response time and power consumption. Finally, a range of client applications that exploit the middleware on different platforms are demonstrated, to prove its usability and effectiveness in enabling, in this scenario, energy monitoring and savings.
Uploads
Papers by I. Vlahavas