Dimitris Kalles
Hellenic Open University, School of Sciences & Technology, Faculty Member
This paper proposes a method to improve ID5R, an incremental TDIDT algorithm. The new method evaluates the quality of attributes selected at the nodes of a decision tree and estimates a minimum number of steps for which these attributes... more
This paper proposes a method to improve ID5R, an incremental TDIDT algorithm. The new method evaluates the quality of attributes selected at the nodes of a decision tree and estimates a minimum number of steps for which these attributes are guaranteed such a selection. This results in reducing overheads during incremental learning. The method is supported by theoretical analysis and experimental results.
Conventional algorithms for decision tree induction use an attribute-value representation scheme for instances. This paper explores the empirical consequences of using set-valued attributes. This simple representational extension, when... more
Conventional algorithms for decision tree induction use an attribute-value representation scheme for instances. This paper explores the empirical consequences of using set-valued attributes. This simple representational extension, when used as a pre-processor for numeric data, is shown to yield significant gains in accuracy combined with attractive build times. It is also shown to improve the accuracy for the second best classification option, which has valuable ramifications for post-processing.
We investigate systematically the impact of human intervention in the training of computer players in a strategy board game. In that game, computer players utilise reinforcement learning with neural networks for evolving their playing... more
We investigate systematically the impact of human intervention in the training of computer players in a strategy board game. In that game, computer players utilise reinforcement learning with neural networks for evolving their playing strategies and demonstrate a slow learning speed. Human intervention can significantly enhance learning performance, but carrying it out systematically seems to be more of a problem of an integrated game development environment as opposed to automatic evolutionary learning.
Summary A new energy-based theory for seismic failure of reinforced concrete tunnel linings is proposed. The above theory has been confirmed by measurements on tunnel lining samples and will be applied to tunnels using an intelligent... more
Summary A new energy-based theory for seismic failure of reinforced concrete tunnel linings is proposed. The above theory has been confirmed by measurements on tunnel lining samples and will be applied to tunnels using an intelligent seismic monitoring system that is under development. This system will detect the deformations in reinforced concrete tunnels in real time and will assess the remaining capacity of the lining to dissipate energy and the degree of damage.
In this paper we offer a report on a university-level programming laboratory course that has been designed on top of a programming library. The course enforces soft skills, such as code inspection and team working, sharpens implementation... more
In this paper we offer a report on a university-level programming laboratory course that has been designed on top of a programming library. The course enforces soft skills, such as code inspection and team working, sharpens implementation skills and creates a bridge between introductory, language-specific instruction and senior-year full-blown programming projects that are usually large but not attending to soft skills. Quite as importantly, it has also delivered a working research tool.
Abstract. We use decision trees and genetic algorithms to analyze the academic performance of students and the homogeneity of tutoring teams in the undergraduate program on Informatics at the Hellenic Open University (HOU). Based on the... more
Abstract. We use decision trees and genetic algorithms to analyze the academic performance of students and the homogeneity of tutoring teams in the undergraduate program on Informatics at the Hellenic Open University (HOU). Based on the accuracy of the generated rules, we examine the applicability of the techniques at large and reflect on how one can deploy such techniques in academic performance alert systems.
UtilNets is a decision-support system (DSS) for rehabilitation planning and optimisation of the maintenance of underground pipe networks of water utilities. The DSS performs reliability based life predictions of the pipes and determines... more
UtilNets is a decision-support system (DSS) for rehabilitation planning and optimisation of the maintenance of underground pipe networks of water utilities. The DSS performs reliability based life predictions of the pipes and determines the consequences of maintenance and neglect over time in order to optimise rehabilitation policy.
Many students who enrol in the undergraduate program on informatics at the Hellenic Open University (HOU) fail the introductory course exams and drop out. We analyze their academic performance, derive short rules that explain success or... more
Many students who enrol in the undergraduate program on informatics at the Hellenic Open University (HOU) fail the introductory course exams and drop out. We analyze their academic performance, derive short rules that explain success or failure in the exams and use the accuracy of these rules to reflect on specific tutoring practices that could enhance success.
Students that enroll in the undergraduate program on informatics at the Hellenic Open University (HOU) demonstrate significant difficulty in advancing beyond the introductory course. We have embarked in an effort to analyze their academic... more
Students that enroll in the undergraduate program on informatics at the Hellenic Open University (HOU) demonstrate significant difficulty in advancing beyond the introductory course. We have embarked in an effort to analyze their academic performance throughout the academic year, as measured by homework assignments, and attempt to derive short rules that explain and predict success or failure in the final exams.
This work deals with stability in incremental induction of decision trees. Stability problems arise when an induction algorithm must revise a decision tree very often and oscillations between similar concepts decrease learning speed. We... more
This work deals with stability in incremental induction of decision trees. Stability problems arise when an induction algorithm must revise a decision tree very often and oscillations between similar concepts decrease learning speed. We introduce a heuristic and an algorithm with theoretical and experimental backing to tackle this problem.
Genetic Algorithms (GAS) have been widely used as an effective search technique. Rather than search from general-to-specific or from simple-to-complex hypotheses, GAS generate successor hypotheses by repeatedly mutating and recombining... more
Genetic Algorithms (GAS) have been widely used as an effective search technique. Rather than search from general-to-specific or from simple-to-complex hypotheses, GAS generate successor hypotheses by repeatedly mutating and recombining parts of the best currently known hypotheses. Decision tree induction has been successfully applied to a broad range of tasks from learning to diagnose medical cases to learning to assess credit risk of loan applicants. The construction of optimal decision trees has been proven to be NP-complete [l].
Abstract In this paper a new algorithm is proposed for skeletonizing binary digital images. The algorithm does not employ the conventional pixel-or non-pixel-based techniques. Instead, it identifies regions of particular shapes in the... more
Abstract In this paper a new algorithm is proposed for skeletonizing binary digital images. The algorithm does not employ the conventional pixel-or non-pixel-based techniques. Instead, it identifies regions of particular shapes in the image and substitutes apppropriate skeleton patterns for them. Initially, as many horizontal and vertical strips as possible are detected. These correspond to rather straight, long and narrow regions in the original image. Any remaining regions correspond to joints between strips.
Abstract We explore the use of genetic algorithms to directly evolve classification decision trees. We argue on the suitability of such a concept learner due to its ability to efficiently search complex hypotheses spaces and discover... more
Abstract We explore the use of genetic algorithms to directly evolve classification decision trees. We argue on the suitability of such a concept learner due to its ability to efficiently search complex hypotheses spaces and discover conditionally dependent as well as irrelevant attributes. The performance of the system is measured on a set of artificial and standard discretized concept-learning problems and compared with the performance of two known algorithms (C4. 5, OneR).
ABSTRACT In this paper we elaborate on the application of reinforcement learning to the design of a new strategy game. We deal with playability and learning issues, attempting to use intelligently generated self-playing sequences to... more
ABSTRACT In this paper we elaborate on the application of reinforcement learning to the design of a new strategy game. We deal with playability and learning issues, attempting to use intelligently generated self-playing sequences to determine playability of various initial board configurations. The machine's a priori knowledge about the game is restricted to the rules only, so, the initially encouraging and intuitive results suggest that this design verification strategy may be useful to a board range of design problems.
Let n atomic players be routing their unsplitable flow on m resources. When each player has the option to drop her current resource and select a better one, and this option is exercised sequentially and unilaterally, then a Nash... more
Let n atomic players be routing their unsplitable flow on m resources. When each player has the option to drop her current resource and select a better one, and this option is exercised sequentially and unilaterally, then a Nash Equilibrium (NE) will be eventually reached. Acting sequentially, however, is unrealistic in large systems.
Abstract. All students of the Hellenic Open University (HOU) attend undergraduate and postgraduate courses at a distance. The lack of a live academic community is reported by many as a drawback in their studies. Systematic exploitation of... more
Abstract. All students of the Hellenic Open University (HOU) attend undergraduate and postgraduate courses at a distance. The lack of a live academic community is reported by many as a drawback in their studies. Systematic exploitation of new communication and collaboration technologies is desirable in the HOU but cannot be imposed universally as the average student's IT competence level is relatively low.
This paper reports on the development of a library of decision tree algorithms in Java. The basic model of a decision tree algorithm is presented and then used to justify the design choices and system architecture issues. The library has... more
This paper reports on the development of a library of decision tree algorithms in Java. The basic model of a decision tree algorithm is presented and then used to justify the design choices and system architecture issues. The library has been designed for flexibility and adaptability. Its basic goal was an open system that could easily embody parts of different conventional as well as new algorithms, without the need of knowing the inner organization of the system in detail.
ABSTRACT This paper presents a machine learning approach to the problems of part-of-speech disambiguation and unknown word guessing, as they appear in Modern Greek. Both problems are cast as classification tasks carried out by decision... more
ABSTRACT This paper presents a machine learning approach to the problems of part-of-speech disambiguation and unknown word guessing, as they appear in Modern Greek. Both problems are cast as classification tasks carried out by decision trees. The data model acquired is capable of capturing the idiosyncratic behavior of underlying linguistic phenomena. Decision trees are induced with three algorithms; the first two produce generalized trees, while the third produces binary trees.
Abstract. We investigate systematically the impact of a minimax tutor in the training of computer players in a strategy board game. In that game, computer players utilise reinforcement learning with neural networks for evolving their... more
Abstract. We investigate systematically the impact of a minimax tutor in the training of computer players in a strategy board game. In that game, computer players utilise reinforcement learning with neural networks for evolving their playing strategies.
Abstract When genetic algorithms are used to evolve decision trees, key tree quality parameters can be recursively computed and re-used across generations of partially similar decision trees. Simply storing instance indices at leaves is... more
Abstract When genetic algorithms are used to evolve decision trees, key tree quality parameters can be recursively computed and re-used across generations of partially similar decision trees. Simply storing instance indices at leaves is sufficient for fitness to be piecewise computed in a lossless fashion. We show the derivation of the (substantial) expected speedup on two bounding case problems and trace the attractive property of lossless fitness inheritance to the divide-and-conquer nature of decision trees.
ABSTRACT There is currently a high activity in the transportation tunnelling industry in the countries of the Union with the highest seismicity.
Abstract. We use decision trees and genetic algorithms to analyze the academic performance of students throughout an academic year at a distance learning university. Based on the accuracy of the generated rules, and on crossexaminations... more
Abstract. We use decision trees and genetic algorithms to analyze the academic performance of students throughout an academic year at a distance learning university. Based on the accuracy of the generated rules, and on crossexaminations of various groups of the same student population, we surprisingly observe that students' performance is clustered around tutors.
The Hellenic Open University has embarked on a large-scale effort to enhance its textbook-based material with content that demonstrably supports the basic tenets of distance learning. The challenge is to set up a framework that allows for... more
The Hellenic Open University has embarked on a large-scale effort to enhance its textbook-based material with content that demonstrably supports the basic tenets of distance learning. The challenge is to set up a framework that allows for the production-level creation, distribution and consumption of content, and at the same time, evaluate the effort in terms of technological, educational and organizational knowledge gained.
ABSTRACT Conventional predictive maintenance involves continuous processing of real-time data from plant sensors of critical variables that are indicators of the health of the equipment. Some intelligent monitoring systems using rules... more
ABSTRACT Conventional predictive maintenance involves continuous processing of real-time data from plant sensors of critical variables that are indicators of the health of the equipment. Some intelligent monitoring systems using rules elicited from maintenance personnel have being developed to infer the causes of impending faults. In this paper we propose a novel approach to intelligent predictive maintenance based on reinforcement learning.
Abstract Two cases of what the author considers grave misconduct in journal reviewing led him to consider how we could improve how journals review submissions. He wanted to treat anonymous peer reviewing as a given because no reasonable... more
Abstract Two cases of what the author considers grave misconduct in journal reviewing led him to consider how we could improve how journals review submissions. He wanted to treat anonymous peer reviewing as a given because no reasonable reengineering of the review process seems to have proposed a workable alternative. Although all the author's data derives from personal experience, the sample is not small, amounting to around 30 rejected submissions to journals and conferences.
Abstract We describe a web intelligent system that demonstrates many of the challenges and implications arising from building such systems on the web. The particular system is an on-line fun portal that adopts several key ideas spanning... more
Abstract We describe a web intelligent system that demonstrates many of the challenges and implications arising from building such systems on the web. The particular system is an on-line fun portal that adopts several key ideas spanning from site design to user walk-through and user-input analysis, in order to “intelligently” adapt to user interests and, consequently, improve user experience.
Abstract In the present communication an approach to the performance optimization of an H. 261 implementation is described. Discrete Cosine Transform (DCT) coefficient pruning is employed and its effect on motion estimation strategies is... more
Abstract In the present communication an approach to the performance optimization of an H. 261 implementation is described. Discrete Cosine Transform (DCT) coefficient pruning is employed and its effect on motion estimation strategies is studied. This interplay produces some very interesting experimental results on the trade-off between quality and speed of video coding
Abstract. Distance learning institutions need to find a way to transplant the benefits of conventional tutoring practices into the development of digital content that is conducive to students' learning needs. Therein lie two great... more
Abstract. Distance learning institutions need to find a way to transplant the benefits of conventional tutoring practices into the development of digital content that is conducive to students' learning needs. Therein lie two great challenges: promote real distance learning effectively and, at the same time, try to accommodate the ability of humans to learn via collaboration.
ABSTRACT This paper presents an electronic auction system as a case study of a distributed computing application using the Java language and the Voyager distributed programming platform. It stresses the importance of state-of-the-art WWW... more
ABSTRACT This paper presents an electronic auction system as a case study of a distributed computing application using the Java language and the Voyager distributed programming platform. It stresses the importance of state-of-the-art WWW development tools that address the run-to-the-market need of electronic commerce applications. Conceptual simplicity and efficient design and implementation are the major advantages from the adoption of a platform where object mobility and persistence are transparently available.
Abstract This paper proposes a method for modelling the quality aspects of e-commerce systems based on Bayesian Networks. The proposed model can be used as an instrument for the measurement of the quality of such systems. The paper... more
Abstract This paper proposes a method for modelling the quality aspects of e-commerce systems based on Bayesian Networks. The proposed model can be used as an instrument for the measurement of the quality of such systems. The paper presents the theoretical background of the model, as well as practical issues arising from its application. Special emphasis is placed on the model's structure and usage, as well as on the potential to extend it and validate it through further AI techniques.
SUMMARY This paper asserts that to transform mainstream surfers to shoppers, we must focus on reinforcing a notion of widespread service and product availability addressing the subconcious fear that technology is for the select few. We... more
SUMMARY This paper asserts that to transform mainstream surfers to shoppers, we must focus on reinforcing a notion of widespread service and product availability addressing the subconcious fear that technology is for the select few. We present key requirements that our infrastructure has to address and a commercial platform that satisfies them. KEYWORDS: e-commerce, infrastructure, instant commerce, networks.
A decision tree is a graphical representation of a procedure for classifying or evaluating an item of interest. It represents a function that maps each element of a domain to a value from a set of values; this value is typically a... more
A decision tree is a graphical representation of a procedure for classifying or evaluating an item of interest. It represents a function that maps each element of a domain to a value from a set of values; this value is typically a symbolic class label or a numerical value. Decision trees are excellent tools for supporting decisions, when a lot of complex information must be taken into account and the reasoning must be supplied for alternative paths (Mitchell, 1997).
Abstract Conventional algorithms for the induction of decision trees use an attribute-value representation scheme for instances. This paper explores the empirical consequences of using set-valued attributes. This simple representational... more
Abstract Conventional algorithms for the induction of decision trees use an attribute-value representation scheme for instances. This paper explores the empirical consequences of using set-valued attributes. This simple representational extension is shown to yield significant gains in speed and accuracy. To do so, the paper also describes an intuitive and practical version of pre-pruning.
In this article we experiment with a 2-player strategy board game where playing models are developed using reinforcement learning and neural networks. The models are developed to speed up automatic game development based on human... more
In this article we experiment with a 2-player strategy board game where playing models are developed using reinforcement learning and neural networks. The models are developed to speed up automatic game development based on human involvement at varying levels of sophistication and density when compared to fully autonomous playing.
Students who enrol in the undergraduate program on informatics at the Hellenic Open University (HOU) demonstrate significant difficulties in advancing beyond the introductory courses. We use decision trees and genetic algorithms to... more
Students who enrol in the undergraduate program on informatics at the Hellenic Open University (HOU) demonstrate significant difficulties in advancing beyond the introductory courses. We use decision trees and genetic algorithms to analyze their academic performance throughout an academic year. Based on the accuracy of the generated rules, we analyze the educational impact of specific tutoring practices.
Abstract Reinforcement learning is considered as one of the most suitable and prominent methods for solving game problems due to its capability to discover good strategies by extended self-training and limited initial knowledge. In this... more
Abstract Reinforcement learning is considered as one of the most suitable and prominent methods for solving game problems due to its capability to discover good strategies by extended self-training and limited initial knowledge. In this paper we elaborate on using reinforcement learning for verifying game designs and playing strategies. Specifically, we examine a new strategy game that has been trained on self-playing games and analyze the game performance after human interaction.
This paper reviews an experiment in human-computer interaction, where interaction takes place when humans attempt to teach a computer to play a strategy board game. We show that while individually learned models can be shown to improve... more
This paper reviews an experiment in human-computer interaction, where interaction takes place when humans attempt to teach a computer to play a strategy board game. We show that while individually learned models can be shown to improve the playing performance of the computer, their straightforward composition results in diluting what was earlier learned. This observation suggests that interaction cannot be easily distributed when one hopes to harness multiple human experts to develop a quality computer player.
Abstract: We investigate systematically the impact of human intervention in the training of computer players in a strategy board game. In that game, computer players utilise reinforcement learning with neural networks for evolving their... more
Abstract: We investigate systematically the impact of human intervention in the training of computer players in a strategy board game. In that game, computer players utilise reinforcement learning with neural networks for evolving their playing strategies and demonstrate a slow learning speed. Human intervention can significantly enhance learning performance, but carry-ing it out systematically seems to be more of a problem of an integrated game development environment as opposed to automatic evolutionary learning.
Abstract: In this paper we examine sorting on the assumption that we do not know in advance which way to sort a sequence of numbers and we set at work simple local comparison and swap operators whose repeating application ends up in... more
Abstract: In this paper we examine sorting on the assumption that we do not know in advance which way to sort a sequence of numbers and we set at work simple local comparison and swap operators whose repeating application ends up in sorted sequences. These are the basic elements of Emerge-Sort, our approach to self-organizing sorting, which we then validate experimentally across a range of samples.
ABSTRACT Distance learning universities usually afford their students the flexibility to advance their studies at their own pace. This can lead to a considerable fluctuation of student populations within a programme's courses. The... more
ABSTRACT Distance learning universities usually afford their students the flexibility to advance their studies at their own pace. This can lead to a considerable fluctuation of student populations within a programme's courses. The evolution of the student population may be an important factor in determining the academic viability of a programme as well as the resources that have to be budgeted and administered.
Distance learning universities usually afford their students the flexibility to advance their studies at their own pace. This can lead to a considerable fluctuation of student populations within a program's courses, possibly affecting the... more
Distance learning universities usually afford their students the flexibility to advance their studies at their own pace. This can lead to a considerable fluctuation of student populations within a program's courses, possibly affecting the academic viability of a program as well as the related required resources. Providing a method that estimates this population could be of substantial help to university management and academic personnel.