Analogical reasoning is considered as one of the main mechanisms underlying creativity. "Thinking out of the box" allows the paradigm shift essential to a creative process. More common is the concept of analogical proportion ("2 is to 4 as 4 is to 8") which can be described within an algebraic framework. When it comes to concepts ("engine is to the car as heart is to the human"), we need to investigate a new way to understand this analogical ratio. In this paper, we take inspiration from the formal framework of information theory for proposing a new approach to the evaluation of analogy between concepts. Using Kolmogorov complexity as a backbone providing a clear semantics, we give a practical interpretation for analogy between words viewed as labeling concepts. Making use of Google as a linguistic resource, we provide an implementation of our definitions: experiments show that the accuracy of our definition is quite acceptable and justify the approach.

Though polarity classification has been extensively explored at document level, there has been little work investigating feature design at sentence level. Due to the small number of words within a sentence, polarity classification at sentence level differs substantially from document-level classification in that resulting bag-of-words feature vectors tend to be very sparse resulting in a lower classification accuracy. In this paper, we show that performance can be improved by adding features specifically designed for sentence-level polarity classification. We consider both explicit polarity information and various linguistic features. A great proportion of the improvement that can be obtained by using polarity information can also be achieved by using a set of simple domain-independent linguistic features.

As robots become more common, it becomes increasingly useful for them to communicate and effectively share knowledge that they have learned through their individual experiences.  Learning from experiences, however, is often-times embodiment-specific; that is, the knowledge learned is grounded in the robot’s unique sensors and actuators.  This type of learning raises questions as to how communication and knowledge exchange via social interaction can occur, as properties of the world can be grounded differently in different robots.  This is especially true when the robots are heterogeneous, with different sensors and perceptual features used to define the properties.  In this paper, we present methods and representations that allow heterogeneous robots to learn grounded property representations, such as that of color categories, and then build models of their similarities and differences in order to map their respective representations.  We use a conceptual space representation, where object properties are learned and represented as regions in a metric space, implemented via supervised learning of Gaussian Mixture Models.  We then propose to use confusion matrices that are built using instances from each robot, obtained in a shared context, in order to learn mappings between the properties of each robot.  Results are demonstrated using two perceptually heterogeneous Pioneer robots, one with a web camera and another with a camcorder.

The paper concentrates on obtaining hidden relationships among individual clauses of complex sentences from the Prague Dependency Treebank. The treebank contains only an information about mutual relationships among individual tokens (words, punctuation marks), not about more complex units (clauses). For the experiments with clauses and their parts (segments) it was therefore necessary to develop an automatic method transforming the original annotation into a scheme describing the syntactic relationships between clauses. The task was complicated by a certain degree of inconsistency in original annotation with regard to clauses and their structure. The paper describes the algorithm of deriving clause-related information from the existing annotation and its evaluation.

Emotions are a ubiquitous component of motivation and learning. We have developed an affective behavior model for intelligent tutoring systems that considers both the affective and knowledge state of the student to generate tutorial actions. The affective behavior model (ABM) was designed based on teachers' expertise obtained through interviews. It relies on a dynamic decision network with a utility measure on both student learning and affect to generate tutorial actions aimed at balancing the two. We have integrated and evaluated the ABM in an educational game to learn number factorization. We carried out a controlled user study to evaluate the impact of the affective model on learning. The results show that for the younger students there is a significant improvement on learning when the affective behavior model is incorporated.

Conference and special track committees responsible for the 2009 FLAIRS conference.

Manual knowledge acquisition is usually a costly and time-consuming process. Automatic knowledge acquisition methods can then significantly support the knowledge engineer. In this paper, we propose an approach for rapid knowledge capture. The methodology is based on textual subgroup mining in order to discover dependencies for rule prototyping.

The main task of part-of-speech (PoS) tagging is to assign the appropriate morphosyntactic category to each word in a sentence. A combination of different PoS taggers usually results in higher tagging accuracy than obtained by the use of only a single tagger. We present a new language and tagset independent system, CombiTagger, which combines automatically the output of several taggers. The system, which is open source, provides algorithms for simple and weighted voting, but it is extensible so that other combination algorithms can be added easily. We demonstrate the functionality of CombiTagger by using it to develop and evaluate combined taggers for Icelandic. The most accurate individual tagger obtains an accuracy of 91.83%. CombiTagger achieves 93.09%-93.41% accuracy by combining the output of five or six taggers using simple and weighted voting.

This paper introduces an information theoretic approach to verification of modular causal probabilistic models. We assume systems which are gradually extended by adding new functional modules, each having a limited domain knowledge captured by a local Bayesian network. Different modules originate from independent design processes. We assume that the local models are correct, which, however does not guarantee globally coherent inference in composed systems. The introduced method supports discovery of significant inter module dependencies which are ignored in the assembled Bayesian network.

Generalized cases are cases that cover a subspace rather than a point in the problem-solution space. Attribute dependent generalized cases are a subclass of generalized cases, which cause a high computational complexity during similarity assessment. We present a new approach for an efficient index-based retrieval of such generalized cases by an improved kd-tree approach. The experimental evaluation demonstrates a significant improvement in retrieval efficiency compared to previous methods.