International Journal on Artificial Intelligence Tools

In this paper we propose a general approach for reasoning in space. The approach is composed of a set of two general constraints to govern the spatial relationships between objects in space, and two rules to propagate relationships between those objects. The approach is based on a novel representation of the topology of the space as a connected set of components using a structure called adjacency matrix which can capture the topology of objects of different complexity in any space dimension. The formalism is used to explain spatial compositions resulting in indefinite and definite relations and it is shown to be applicable to reasoning in the temporal domain. The main contribution of the formalism is that it provides means for constructing composition tables for objects with arbitrary complexity in any space dimension. A new composition table between spatial objects of different types is presented. A major advantage of the method is that reasoning between objects of any complexity can be achieved in a defined limited number of steps. Hence, the incorporation of spatial reasoning mechanisms in spatial information systems becomes possible.

Human beings reason about different aspects of space (such as relative orientation, cardinal directions, distance, size and shape of objects) quite easily. With the aim of simulating human behavior, several models for these spatial concepts have been developed in the recent years. Cognitive considerations have made these frameworks qualitative, because they seem to deal better with the imprecision that human perception provides. However, an operational model to reason with all these spatial aspects in an integrated way has not been developed, up to now. The first aim of our research work has been the integration of different spatial concepts into the same spatial model which has been accomplished thanks to the definition of an operational model based on Constrain Logic Programming extended with Constraint Handling Rules. Although other aspects of space have been successfully represented by these techniques [2], in this paper we focus our attention in positional information, that is, orientation integrated with distance information. The Constraint Solver developed for managing positional information has a temporal complexity of O(n)³, where n is the number of spatial landmarks considered in the reasoning process. The second aim of our work is to apply qualitative spatial reasoning to develop a Qualitative Navigation Simulator.

A system is described that uses a mixed-level knowledge representation based on standard Horn Clause Logic to represent (part of) the meaning of natural language documents. A variable-depth search strategy is outlined that distinguishes between the different levels of abstraction in the knowledge representation to locate specific passages in the documents. A detailed description of the linguistic aspects of the system is given. Mixed-level representations as well as variable-depth search strategies are applicable in fields outside that of NLP.

Classification involves associating instances with particular classes by maximizing intra-class similarities and minimizing inter-class similarities. Thus, the way similarity among instances is measured is crucial for the success of the system. In case-based reasoning, it is assumed that similar problems have similar solutions. The case-based approach to classification is founded on retrieving cases from the case base that are similar to a given problem, and associating the problem with the class containing the most similar cases.

Similarity-based retrieval tools can advantageously be used in building flexible retrieval and classification systems. Case-based classification uses previously classified instances to label unknown instances with proper classes. Classification accuracy is affected by the retrieval process – the more relevant the instances used for classification, the greater the accuracy.

The paper presents a novel approach to case-based classification. The algorithm is based on a notion of similarity assessment and was developed for supporting flexible retrieval of relevant information. Case similarity is assessed with respect to a given context that defines constraints for matching. Context relaxation and restriction is used for controlling the classification accuracy. The validity of the proposed approach is tested on real-world domains, and the system's performance, in terms of accuracy and scalability, is compared to that of other machine learning algorithms.

Data mining algorithms including maching learning, statistical analysis, and pattern recognition techniques can greatly improve our understanding of data warehouses that are now becoming more widespread. In this paper, we focus on classification algorithms and review the need for multiple classification algorithms. We describe a system called , which was designed to help choose the appropriate classification algorithm for a given dataset by making it easy to compare the utility of different algorithms on a specific dataset of interest. not only provides a workbench for such comparisons, but also provides a library of C++ classes to aid in the development of new algorithms, especially hybrid algorithms and multi-strategy algorithms. Such algorithms are generally hard to code from scratch. We discuss design issues, interfaces to other programs, and visualization of the resulting classifiers.

The Processor Configuration Problem (PCP) is a real life Constraint Optimization Problem. The task is to link up a finite set of processors into a network, whilst minimizing the maximum distance between these processors. Since each processor has a limited number of communication channels, a carefully planned layout will help reduce the overhead for message switching.

In this paper, we present a Genetic Algorithm (GA) approach to the PCP. Our technique uses a mutation-based GA, a function that produces schemata by analyzing previous solutions, and an efficient data representation. Our approach has been shown to out-perform other published techniques in this problem.

In many applications, such as decision support, negotiation, planning, scheduling, etc., one needs to express requirements that can be only partially satisfied. In order to express such requirements, we propose a technique called forward-tracking. Intuitively, forward-tracking is a kind of dual of chronological back-tracking: if a program globally fails to find a solution, then a new execution is started from a program point and a state 'forward' in the computation tree. This search technique is applied to the specific paradigm of constraint logic programming, obtaining a powerful extension that preserves all the useful properties of the original scheme. We report on the successful practical application of forward-tracking to the evolutionary training of (constrained) neural networks, and discuss other interesting applications.

A lot of recent research papers focus on the challenging problem of the combination of genetic algorithms and artificial neural networks. Developmental and molecular biology may be a source of inspiration for designing powerful artificial neurogenesis systems allowing the generation of complex modular structures. This paper describes in details such a neurogenesis model associated with an evolutionary process and its applications to the cart-pole problem and to the control of a mobile robot. Early results demonstrate the surprising efficiency of this methodology and give hints to continue the research towards the generation of more complex adaptive neural networks.

The symbolic level of a dynamic scene interpretation system is presented. This symbolic level is based on plan prototypes represented by Petri nets whose interpretation is expressed thanks to 1st order cubes, and on a reasoning aiming at instantiating the plan prototypes with objects delivered by the numerical processing of sensor data. A purely symbolic meta-structure, grounded on the lattice theory, is then proposed to deal with the symbolic uncertainty issues. Examples on real world data are given.

Many real-world problems involve measures of objectives that may be dynamically optimized. The application of evolutionary algorithms, such as genetic algorithms, in time dependent optimization is currently receiving growing interest as potential applications are numerous ranging from mobile robotics to real time process command. Moreover, constant evaluation functions skew results relative to natural evolution so that it has become a promising gap to combine effectiveness and diversity in a genetic algorithm. This paper features both theoretical and empirical analysis of the behavior of genetic algorithms in such an environment. We present a comparison between the effectivenss of traditional genetic algorithm and the dual genetic algorithm which has revealed to be a particularly adaptive tool for optimizing a lot of diversified classes of functions. This comparison has been performed on a model of dynamical environments which characteristics are analyzed in order to establish the basis of a testbed for further experiments. We also discuss fundamental properties that explain the effectiveness of the dual paradigm to manage dynamical environments.