El Bayesianismo y la Justificación de la Inducción
- Pinto, Sílvio
The appearance of Bayesian inductive logic has prompted a renewed optimism about the possibility of justification of inductive rules The justifying argument for the 'rides of such a logic is the famous Dutch Book Argument (Ramsey-de Finetti’s theorem) The issue winch divides the theoreticians of induction concerns the question of whether this argument can indeed legitimize Bayesian conditionalization rides Here I will be firstly interested in showing that the Ramsey de Finetti's argument cannot establish that the use of the mentioned conditionalization rides is the best option against Dutch Book betting strategies except in special circum stances I suggest secondly...
Grammar Learning Using Inductive Logic Programming
- Stephen Pulman; James Cussens
This paper gives a brief introduction to a particular machine learning method known as inductive logic programming. It is argued that this method, unlike many current statistically based machine learning methods, implies a view of grammar learning that bears close affinity to the views linguists have of the `logical problem of language acquisition'. Two experiments
Methodological levels of abductive logic and its application in analyzing knowledge classification systems
- Mohammad Khandan; Gholamreza Fadaei; Mohammad Reza Vasfi
Purpose: This paper distinguishes between “methodology” and “method” and discusses methodological levels of abductive logic as related to the knowledge classification systems. The purpose is to clarify the application of abductive logic for analyzing knowledge classification systems as an alternative for mainstream logics in the field, i.e. inductive and deductive logic.
Methodology: Conceptual analysis.
Findings: Abductive logic approaches reality as a social construction. It is obviously an interpretative and constructivist approach. Therefore its aim is to understand the social phenomena that socially constructed by social agents. Therefore any given knowledge classification system is a socially constructed phenomenon reflecting subjective biases and prejudices...
Replicability of Experiment
- John D. Norton
The replicability of experiment is routinely offered as the gold standard of evidence. I argue that it is not supported by a universal principle of replicability in inductive logic. A failure of replication may not impugn a credible experimental result; and a successful replication can fail to vindicate an incredible experimental result. Rather, employing a material approach to inductive inference, the evidential import of successful replication of an experiment is determined by the prevailing background facts. Commonly, these background facts do support successful replication as a good evidential guide and this has fostered the illusion of a deeper, exceptionless principle.
Learning Relational Event Models from Video
- Krishna S. R. Dubba; Anthony G. Cohn; David C. Hogg; Mehul Bhatt; Frank Dylla
Event models obtained automatically from video can be used in applications ranging from abnormal event detection to content based video retrieval. When multiple agents are involved in the events, characterizing events naturally suggests encoding interactions as relations. Learning event models from this kind of relational spatio-temporal data using relational learning techniques such as Inductive Logic Programming (ILP) hold promise, but have not been successfully applied to very large datasets which result from video data. In this paper, we present a novel framework remind (Relational Event Model INDuction) for supervised relational learning of event models from large video datasets using ILP....
A Delta Debugger for ILP Query Execution
- Remko Tronçon; Bart Demoen; Gerda Janssens
Abstract. Because query execution is the most crucial part of Inductive Logic Programming (ILP) algorithms, a lot of effort is invested in developing faster execution mechanisms. These execution mechanisms typically have a low-level implementation, making them hard to debug. Moreover, other factors such as the complexity of the problems handled by ILP algorithms and size of the code base of ILP data mining systems make debugging at this level a very difficult job. In this work, we present the trace-based debugging approach currently used in the development of new execution mechanisms in hipP, the engine underlying the ACE Data Mining...
Analysis of heuristic rule evaluation measures
- Filip Železný; Nada Lavra
Numerous methods are used for performance evaluation in machine learning and knowledge discovery. In inductive logic programming (ILP)- an important subfield of KDD- where the goal is the automatic construction of hypothesis (defined in a formal language) on the basis of an extensional set of facts (examples) and certain background knowledge, various measures are used to select the most promising element (clause) to be part of the current hypothesis. In the presence of noise (errors) in the input data, we usually trade off between the clause’s accuracy and generality. The trade-off is expressed by a clause-quality measure (heuristic). Recently, some...
Bottom-Up Learning of Logic Programs for Information Extraction from Hypertext Documents
- Bernd Thomas
We present an inductive logic programming bottom-up learning algorithm (BFOIL) for synthesizing logic programs for multi-slot information extraction from hypertext documents. BFOIL learns from positive examples only. Furthermore we introduce a logical and relational based representation for hypertext documents (TDOM). We briefly discuss several BFOIL refinements and show very promising results of our system LIPX in comparison to state of the art IE systems.
Margin-Based First-Order Rule Learning
- Ulrich Rückert; Stefan Kramer
Learning sets of first-order rules has a long tradition in machine learning and inductive logic programming. While most traditional systems follow a separateand-conquer approach, many modern systems are based on statistical considerations, such as ensemble theory, large margin classification or graphical models.
Department: Computational Logic
- Georg Rammé; Supervisors Prof; Steffen Hölldobler; Technische Universität Dresden
Since the early days of theorem proving and inductive logic programming several θ-subsumption algorithms have been developed. Recently, the focus came back to θ-subsumption due to its relevance in planning within first-order Markov Decision Processes. More than one formalism has been adopted to describe the language and the algorithms. Many experimental evaluations have been performed, but all focusing only on some algorithms and a particular domain. In this thesis, we will present the most popular θ-subsumption algorithms within a unified framework for a fair comparison. Further, we will describe the domains in which θsubsumption is used and present a huge...
Representing, learning, and executing operational concepts
- Volker Klingspor; Stefan Sklorz
Abstract. On the one hand side operational concepts enable the user to get information about what a robot has been done and on the other hand side, they enable the user to control the robot. In this way, they function as elements of a high level command language, easy to understand by human users. Operational concepts should be general enough to be applied in di erent but similar environments like o ce rooms even if the environment is totally unknown. They combine sensing and action of a mobile robot situated in the real world without the need of additional environmental...
Learning concepts from sensor data of a mobile robot
- Katharina J. Morik; Anke D. Rieger
Abstract. Machine learning can be a most valuable tool for improving the exibility and e ciency of robot applications. Many approaches to applying machine learning to robotics are known. Some approaches enhance the robot's high-level processing, the planning capabilities. Other approaches enhance the low-level processing, the control of basic actions. In contrast, the approach presented in this paper uses machine learning for enhancing the link between the low-level representations of sensing and action and the high-level representation of planning. The aim is to facilitate the communication between the robot and the human user. A hierarchy of concepts is learned from...
An Efficient Approximation to Lookahead in Relational Learners
- Jan Struyf; Jesse Davis
Abstract. Greedy machine learning algorithms suffer from shortsightedness, potentially returning suboptimal models due to limited exploration of the search space. Greedy search misses useful refinements that yield a significant gain only in conjunction with other conditions. Relational learners, such as inductive logic programming algorithms, are especially susceptible to this problem. Lookahead helps greedy search overcome myopia; unfortunately it causes an exponential increase in execution time. Furthermore, it may lead to overfitting. We propose a heuristic for greedy relational learning algorithms that can be seen as an efficient, limited form of lookahead. Our experimental evaluation shows that the proposed heuristic yields...
Acquiring and adapting probabilistic models of agent conversation
- Felix Fischer; Michael Rovatsos; Gerhard Weiss
Communication in multiagent systems (MASs) is usually governed by agent communication languages (ACLs) and communication protocols carrying a clear cut semantics. With an increasing degree of openness, however, the need arises for more flexible models of communication that can handle the uncertainty associated with the fact that adherence to a supposedly agreed specification of possible conversations cannot be ensured on the side of other agents. As one example for such a model, interaction frames follow an empirical semantics view of communication, where meaning is defined in terms of expected consequences, and allow for a combination of existing expectations with empirical...
Automatic inference of indexing rules for MEDLINE
- Shooshan Sonya E; Névéol Aurélie; Claveau Vincent
Indexing is a crucial step in any information retrieval system. In MEDLINE, a widely used database of the biomedical literature, the indexing process involves the selection of Medical Subject Headings in order to describe the subject matter of articles. The need for automatic tools to assist MEDLINE indexers in this task is growing with the increasing number of publications being added to MEDLINE.
In this paper, we describe the use and the customization of Inductive Logic Programming (ILP) to infer indexing rules that may be used to produce automatic indexing recommendations for MEDLINE indexers.
Reshaping the Science of Reliability with the Entropy Function
- Paolo Rocchi; Giulia Capacci
The present paper revolves around two argument points. As first, we have observed a certain parallel between the reliability of systems and the progressive disorder of thermodynamical systems; and we import the notion of reversibility/irreversibility into the reliability domain. As second, we note that the reliability theory is a very active area of research which although has not yet become a mature discipline. This is due to the majority of researchers who adopt the inductive logic instead of the deductive logic typical of mature scientific sectors. The deductive approach was inaugurated by Gnedenko in the reliability domain. We mean to...
¿Es Posible Aprender Inductivamente de la Experiencia?
- Dayron A. Arboleda-Quintero; Margarita E. Patiño
With the birth of probability calculus as a new risk quantification methodology, it also appears the debate with respect to the viability of inductive probability or to the fact that classic statistics uses inductive or deductive methods to draw its conclusions which is the basis of the scientific method. Such discussion becomes more intense with the strengthening of Bayesian statistics that seems to be the most promissory answer in favor of inductive learning and which is supported here by the authors of this article. There are many philosophers that reject the possibility to learn inductively from the experience. The most...
Hedging Uncertainty in Rough Set-based Approach with Fuzzy Decision
- Ning Chen; Bo-Qin Feng; Haixiao Wang; Hao Zhang
This study introduced a technique for authenticating the vehicle engines by comparing the images of the imprints of the identification number acquired when the vehicle was first registered and the ones acquired from the routine yearly vehicle inspection. The images were taken by rubbing a pencil over a piece of paper covered over the images and then scanned into a computer as binary image. Due to the nature of the acquiring technique, the acquired images have lots of artifacts caused by the shape and the condition of the engine surface and unevenness of rubbing the pencils by hand. The rough...
La transición y el proceso de adaptación en la Educación Superior: un estudio con estudiantes de una escuela de enfermería y de una escuela de educación
- Rita Sousa; Amélia Lopes; Elisabete Ferreira
La entrada en la Educación Superior es un periodo de transición en la vida del estudiante por lo que la relación que el estudiante crea con la propia institución educativa y la forma en la que se implica en sus dinámicas es primordial para su buena adaptación y éxito educativo. En este sentido, resulta relevante abordar la integración de los estudiantes en grupos en los que desarrollen un sentimiento de pertenencia. En el contexto de la Educación Superior, se puede considerar en este caso la Praxis Académica y la Asociación de Estudiantes, instituciones que de forma general más movilizan a...
Inférence de règles de propagation syntaxique pour l’alignement de mots
- Sylwia Ozdowska; Vincent Claveau
This paper presents and evaluates an original approach to automatically align bitexts at the word level. It relies on a syntactic dependency analysis of the texts and uses a machinelearning technique, namely inductive logic programming, to automatically infer rules called propagation rules. These rules make the most of the syntactic information to precisely align words. This approach is entirely automatic, uses very few training data, and its results rival the ones of the best existing alignment systems. Moreover, syntactic isomorphisms between the two spotted languages are easily identified through the inferred rules.