The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
The ancient Celts believed that there were three objects of intellect: the true, the beautiful, and the beneficial. Like all good ideas, ILP combines aspects all three of these objects. In this talk I focus on how to obtain the benefits of ILP. I describe applications of ILP to drug design, toxicology, protein function prediction, chemical pathway discovery, and automatic scientific discovery. The...
This talk will address the issue of designing architectures for agents that need to be able to adapt to changing circumstances during deployment. From a scientific point of view, the primary challenge is to design agent architectures that seamlessly integrate reasoning and learning capabilities. That this is indeed a challenge is largely due to the fact that reasoning and knowledge representation...
In many data mining tools that support regression tasks, training data are stored in a single table containing both the target field (dependent variable) and the attributes (independent variables). Generally, only intra-tuple relationships between the attributes and the target field are found, while inter-tuple relationships are not considered and (inter-table) relationships between several tuples...
We study several complexity parameters for first-order formulas and their suitability for first order learning models. We show that the standard notion of size is not captured by sets of parameters that are used in the literature. We then identify an alternative notion of size and a simple set of parameters that are useful in this sense. Matching VC-dimension lower bounds complete the picture showing...
We describe an efficient implementation (MRDTL-2) of the Multi-relational decision tree learning (MRDTL) algorithm [23] which in turn was based on a proposal by Knobbe et al. [19]. We describe some simple techniques for speeding up the calculation of sufficient statistics for decision trees and related hypothesis classes from multi-relational data. Because missing values are fairly common in many...
This work presents the application of theory revision to the design of distributed databases to automatically revise a heuristic-based algorithm (called analysis algorithm) through the use of the FORTE system. The analysis algorithm decides the fragmentation technique to be used in each class of the database and its Prolog implementation is provided as the initial domain theory. Fragmentation schemas...
In the case of concept learning from positive and negative examples, it is rarely possible to find a unique discriminating conjunctive rule; in most cases, a disjunctive description is needed. This problem, known as disjunctive learning, is mainly solved by greedy methods, iteratively adding rules until all positive examples are covered. Each rule is determined by discriminating properties, where...
We believe that AI programs written for discovery tasks will need to simultaneously employ a variety of reasoning techniques such as induction, abduction, deduction, calculation and invention. We describe the HR system which performs a novel ILP routine called automated theory formation. This combines inductive and deductive reasoning to form clausal theories consisting of classification rules and...
Efficiency of the first-order logic proof procedure is a major issue when deduction systems are to be used in real environments, both on their own and as a component of larger systems (e.g., learning systems). Hence, the need of techniques that can speed up such a process. This paper proposes a new algorithm for matching first-order logic descriptions under θ-subsumption that is able to return the...
This work aims at improving the scalability of memory usage in Inductive Logic Programming systems. In this context, we propose two efficient data structures: the Trie, used to represent lists and clauses; and the RL-Tree, a novel data structure used to represent the clauses coverage. We evaluate their performance in the April system using well known datasets. Initial results show a substantial reduction...
Relational reinforcement learning is a Q-learning technique for relational state-action spaces. It aims to enable agents to learn how to act in an environment that has no natural representation as a tuple of constants. In this case, the learning algorithm used to approximate the mapping between state-action pairs and their so called Q(uality)-value has to be not only very reliable, but it also has...
In this paper, we investigate condensation of a clause. First, we extend a substitution graph introduced by Scheffer et al. (1996) to a total matcher graph. Then, we give a correct proof of the relationship between subsumption and the existence of cliques in a total matcher graph. Next, we introduce the concept of width of a clique in a total matcher graph. As a corollary of the above relationship,...
Boosting has established itself as a successful technique for decreasing the generalization error of classification learners by basing predictions on ensembles of hypotheses. While previous research has shown that this technique can be made to work efficiently even in the context of multirelational learning by using simple learners and active feature selection, such approaches have relied on simple...
Propositionalization has already been shown to be a promising approach for robustly and effectively handling relational data sets for knowledge discovery. In this paper, we compare up-to-date methods for propositionalization from two main groups: logic-oriented and database-oriented techniques. Experiments using several learning tasks – both ILP benchmarks and tasks from recent international data...
This paper deals with learning in -log, a hybrid language that merges the function-free Horn clause language Datalog and the description logic . Our application context is descriptive data mining. We introduce -queries, a rule-based form of unary conjunctive queries in -log, and a generality order ≽ B for...
In this paper we present and prove both negative and positive theoretical results concerning the representation and evaluation of first-order logic clauses using genetic algorithms. Over the last few years, a few approaches have been proposed aiming to combine genetic and evolutionary computation (EC) with inductive logic programming (ILP). The underlying rationale is that evolutionary algorithms,...
The key ingredient for any distance-based method in machine learning is a proper distance between individuals of the domain. Distances for structured data have been investigated for some time, but no general agreement has been reached. In this paper we use first-order terms for knowledge representation, and the distances introduced are metrics that are defined on the lattice structure of first-order...
Bioinformatics is characterised by a growing diversity of large-scale databases containing information on genetics, proteins, metabolism and disease. It is widely agreed that there is an increasingly urgent need for technologies which can integrate these disparate knowledge sources. In this paper we propose that not only is machine learning a good candidate technology for such data integration, but...
Given a sample from an unknown probability distribution over strings, there exist algorithms for inferring the structure and parameters of stochastic grammatical representations of the unknown distribution, i.e. string grammars. Despite the fact that research on grammatical representations of sets of graphs has been conducted since the late 1960’s, almost no work has considered the possibility of...
Induction of the effects of actions considered here consists in learning an action description of a dynamic system from evidence on its behavior. General logic-based induction methods can deal with this problem but, unfortunately, most of the solutions provided have the frame problem. To cope with the frame problem induction under suitable nonmonotonic formalisms has to be used, though this kind of...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.