141.
Experiments in numerical reasoning with Inductive Logic Programming
- Ashwin Srinivasan,Rui Camacho
Using problem-specific background knowledge, computer programs developed
within the framework of Inductive Logic Programming (ILP) have
been used to construct restricted first-order logic solutions to scientific
problems. However, their approach to the analysis of data with substantial
numerical content has been largely limited to constructing clauses
that: (a) provide qualitative descriptions ("high", "low" etc.) of the values
of response variables; and (b) contain simple inequalities restricting
the ranges of predictor variables. This has precluded the application of
such techniques to scientific and engineering problems requiring a more
sophisticated approach. A number of specialised methods have been suggested
to remedy this. In contrast, we have chosen to take advantage of
the fact...
142.
An Inductive Logic Programming Method for Corpus-based Parser Construction
- John M. Zelle,Raymond J. Mooney
Empirical methods for building natural language systems has become an important area of research
in recent years. Most current approaches are based on propositional learning algorithms and have been
applied to the problem of acquiring broad-coverage parsers for relatively shallow (syntactic) representations.
This paper outlines an alternative empirical approach based on techniques from a subfield of
machine learning known as Inductive Logic Programming (ILP). ILP algorithms, which learn relational
(first-order) rules, are used in a parser acquisition system called Chill that learns rules to control the
behavior of a traditional shift-reduce parser. Using this approach, Chill is able to learn parsers for a variety
of different types...
143.
Inductive Logic Programming With Large-Scale Unstructured Data
- Michael Bain,Ashwin Srinivasan
We report some recent developments from an ongoing project in which
a chess endgame domain is providing benchmark experimental tests for
the study of concept learning. The King and Rook against King (KRK)
endgame is simple enough in chess terms but provides concept learning
tasks which can be demanding, as evidenced in previous studies by a number
of authors. For learning systems these tasks have highlighted problems
of representation, such as the ability to express the structural relationships
to be found in learning examples, and other issues like correctness, compression
and comprehensibility. Our current focus is on Inductive Logic
Programming methods which are based on previously developed systems
for the...
144.
Inductive Constraint Logic and the Mutagenesis Problem
- Wim Van,Laer Hendrik Blockeel,Luc De Raedt
A novel approach to learning first order logic formulae from positive and negative
examples is incorporated in a system named ICL (Inductive Constraint
Logic). In ICL, examples are viewed as interpretations which are true or false
for the target theory, whereas in present inductive logic programming systems,
examples are true and false ground facts (or clauses). Furthermore, ICL uses a
clausal representation, which corresponds to a conjunctive normal form where
each conjunct forms a constraint on positive examples, whereas classical learning
techniques have concentrated on concept representations in disjunctive normal
form.
We present some experiments with this new system on the mutagenesis problem.
These experiments illustrate some of the differences...
145.
Inductive Constraint Logic and the Mutagenesis Problem
- Hendrik Blockeel,Wim Van Laer,Luc De Raedt
A novel approach to learning first order logic formulae from positive and negative examples is incorporated in a system named ICL (Inductive Constraint Logic). In ICL, examples are viewed as interpretations which are true or false for the target theory, whereas in present inductive logic programming systems, examples are true and false ground facts (or clauses). Furthermore, ICL uses a clausal representation, which corresponds to a conjunctive normal form where each conjunct forms a constraint on positive examples, whereas classical learning techniques have concentrated on concept representations in disjunctive normal form. We present some experiments with this new system on...
146.
Inductive Constraint Logic and the Mutagenesis Problem
- Hendrik Blockeel,Wim Van Laer,Luc De Raedt
A novel approach to learning first order logic formulae from positive and
negative examples is incorporated in a system named ICL (Inductive Constraint
Logic). In ICL, examples are viewed as interpretations which are true
or false for the target theory, whereas in present inductive logic programming
systems, examples are true and false ground facts (or clauses). Furthermore,
ICL uses a clausal representation, which corresponds to a conjunctive normal
form where each conjunct forms a constraint on positive examples, whereas
classical learning techniques have concentrated on concept representations in
disjunctive normal form.
We present some experiments with this new system on the mutagenesis
problem. These experiments illustrate some of the differences...
147.
Inductive Constraint Logic and the Mutagenesis Problem
- Hendrik Blockeel,Wim Van Laer,Luc De Raedt
A novel approach to learning first order logic formulae from positive and
negative examples is incorporated in a system named ICL (Inductive Constraint
Logic). In ICL, examples are viewed as interpretations which are true
or false for the target theory, whereas in present inductive logic programming
systems, examples are true and false ground facts (or clauses). Furthermore,
ICL uses a clausal representation, which corresponds to a conjunctive normal
form where each conjunct forms a constraint on positive examples, whereas
classical learning techniques have concentrated on concept representations in
disjunctive normal form.
We present some experiments with this new system on the mutagenesis
problem. These experiments illustrate some of the differences...
148.
Inductive Constraint Logic and the Mutagenesis Problem
- Hendrik Blockeel,Wim Van Laer,Luc De Raedt
A novel approach to learning first order logic formulae from positive and
negative examples is incorporated in a system named ICL (Inductive Constraint
Logic). In ICL, examples are viewed as interpretations which are true
or false for the target theory, whereas in present inductive logic programming
systems, examples are true and false ground facts (or clauses). Furthermore,
ICL uses a clausal representation, which corresponds to a conjunctive normal
form where each conjunct forms a constraint on positive examples, whereas
classical learning techniques have concentrated on concept representations in
disjunctive normal form.
We present some experiments with this new system on the mutagenesis
problem. These experiments illustrate some of the differences...
149.
Inductive Constraint Logic Programming: An Overview
- Srinivas Padmanabhuni,Aditya K. Ghose
. This paper provides a brief introduction and overview of the
emerging area of Inductive Constraint Logic Programming (ICLP). It
discusses some of the existing work in the area and presents some of the
research issues and open questions that need to be addressed.
1 Introduction
Inductive Logic Programming (ILP) refers to a class of machine learning algorithms
where the agent learns a first-order theory from examples and background
knowledge. The ILP framework in machine learning is perhaps the most general
of all because of the complexity of the concepts learned. The use of first-order
logic programs as the underlying representation makes ILP systems more powerful
and useful than the...
150.
Inductive Constraint Logic and the Mutagenesis Problem
- Hendrik Blockeel,Wim Van Laer,Luc De Raedt
A novel approach to learning first order logic formulae from positive and
negative examples is incorporated in a system named ICL (Inductive Constraint
Logic). In ICL, examples are viewed as interpretations which are true
or false for the target theory, whereas in present inductive logic programming
systems, examples are true and false ground facts (or clauses). Furthermore,
ICL uses a clausal representation, which corresponds to a conjunctive normal
form where each conjunct forms a constraint on positive examples, whereas
classical learning techniques have concentrated on concept representations in
disjunctive normal form.
We present some experiments with this new system on the mutagenesis
problem. These experiments illustrate some of the differences...
151.
Pharmacophore Discovery using the Inductive Logic Programming System Progol
- Paul Finn,David Page,Ron Kohavi,Foster Provost
. This paper presents a case study of a machine-aided knowledge discovery process
within the general area of drug design. Within drug design, the particular problem of pharmacophore
discovery is isolated, and the Inductive Logic Programming (ILP) system progol is
applied to the problem of identifying potential pharmacophores for ACE inhibition. The case
study reported in this paper supports four general lessons for machine learning and knowledge
discovery, as well as more specific lessons for pharmacophore discovery, for Inductive Logic Programming,
and for ACE inhibition. The general lessons for machine learning and knowledge
discovery are as follows.
1. An initial rediscovery step is a useful tool when approaching...
152.
Scaling up Inductive Logic Programming: An Evolutionary Wrapper Approach
- Philip G. K,Reiser And,Patricia J. Riddle
Inductive logic programming (ILP) algorithms are classification algorithms that construct classifiers
represented as logic programs. ILP algorithms have a number of attractive features, notably the ability
to make use of declarative background (user-supplied) knowledge. However, ILP algorithms deal poorly
with large data sets (? 10
examples) and their widespread use of the greedy set-covering algorithm
renders them susceptible to local maxima in the space of logic programs.
153.
An Inductive Logic Programming Query Language for Database Mining (Extended Abstract)
)
Luc De Raedt
Department of Computer Science, Katholieke Universiteit Leuven
Celestijnenlaan 200A, B-3001 Heverlee, Belgium
Tel: ++ 32 16 32 76 43 Fax : ++ 32 16 32 79 96
Luc.DeRaedt@cs.kuleuven.ac.be
Abstract. First, a short introduction to inductive logic programming
and machine learning is presented and then an inductive database mining
query language RDM (Relational Database Mining language). RDM
integrates concepts from inductive logic programming, constraint logic
programming, deductive databases and meta-programming into a flexible
environment for relational knowledge discovery in databases. The
approach is motivated by the view of data mining as a querying process
(see Imielinkski and Mannila, CACM 96). Because the primitives
of the presented query language can easily be...
154.
Pharmacophore Discovery using the Inductive Logic Programming System Progol
- Paul Finn,David Page,Ronny Kohavi,Foster Provost
. This paper is a case study of a machine aided knowledge discovery process within the
general area of drug design. More specifically, the paper describes a sequence of experiments in
which an Inductive Logic Programming(ILP) system is used for pharmacophore discovery. Within
drug design, a pharmacophore is a description of the substructure of a ligand (a small molecule)
which is responsible for medicinal activity. This medicinal activity is produced by interaction
between the ligand and a binding site on a target protein. ILP was chosen by the domain expert
(first author) at Pfizer since active molecules are most naturally described, in relational terms, as
requiring a...
155.
Soccer Agents Learning from Past Behavior with Inductive Logic Programming
- Tohgoroh Matsui,Kazuo Kashiwabara,Nobuhiro Inuzuka,Hirohisa Seki,Hidenori Itoh
this paper, we propose a framework for inductive learning soccer agents (ILSAs), which acquire knowledge from their own past behavior and behave based on the acquired knowledge. Inductive learning is a machine learning framework which is based on generalization of examples. Humanity uses induction to learn a rule of a concept and soccer agents had better use induction as well. We use Inductive Logic Programming (ILP) for soccer agents' inductive learning. ILP combines inductive machine learning and logic programming. Many systems of ILP such as FOIL [3] and Progol [4], can learn complex concepts or rules, which can be expressed...
156.
A Comparative Study of Inductive Logic Programming Methods for Software Fault Prediction
- William W. Cohen,Prem Devanbu
We evaluate inductive logic programming
(ILP) methods for predicting fault density in
C++ classes. In this problem, each training
example is a C++ class definition, represented
as a calling tree, and labeled as "positive
" iff faults (i.e., errors) were discovered
in its implementation. We compare two ILP
systems, FOIL and FLIPPER, and explore
the reasons for their differing performance,
using both natural and artificial data. We
then propose two extensions to FLIPPER:
a user-directed bias towards easy-to-evaluate
clauses, and an extension that allows FLIPPER
to learn "counting clauses". Counting
clauses augment logic programs with a
variation of the "number restrictions" used
in description logics, and significantly improve
performance on this problem when
prior knowledge is used.
1 INTRODUCTION
In...
157.
Sorted Downward Refinement: Building Background Knowledge into a Refinement Operator for Inductive Logic Programming
- Alan M. Frisch
Since its inception, the field of inductive logic programming has been centrally concerned
with the use of background knowledge in induction. Yet, surprisingly, no serious
attempts have been made to account for background knowledge in refinement operators
for clauses, even though such operators are one of the most important, prominent and
widely-used devices in the field. This paper shows how a sort theory, which encodes taxonomic
knowledge, can be built into a downward, subsumption-based refinement operator
for clauses.
Most of this paper was written while the author was a visiting researcher in the Meme Media Laboratory
of the University of Hokkaido, Japan.
y
Email: frisch@cs.york.ac.uk. Phone: +44 1904 432745. Fax:...
158.
A Comparative Study of Inductive Logic Programming Methods for Software Fault Prediction
- William W. Cohen,Prem Devanbu
We evaluate inductive logic programming
(ILP) methods for predicting fault density in
C++ classes. In this problem, each training
example is a C++ class definition, represented
as a calling tree, and labeled as "positive
" iff faults (i.e., errors) were discovered
in its implementation. We compare two ILP
systems, FOIL and FLIPPER, and explore
the reasons for their differing performance,
using both natural and artificial data. We
then propose two extensions to FLIPPER:
a user-directed bias towards easy-to-evaluate
clauses, and an extension that allows FLIPPER
to learn "counting clauses". Counting
clauses augment logic programs with a
variation of the "number restrictions" used
in description logics, and significantly improve
performance on this problem when
prior knowledge is used.
1 INTRODUCTION
In...
159.
Some lower bounds for the Computational Complexity of Inductive Logic Programming
- Jorg-uwe Kietz
The field of Inductive Logic Programming (ILP), which is concerned
with the induction of Horn clauses from examples and background knowledge,
has received increased attention over the last time. Recently, some
positive results concerning the learnability of restricted logic programs
have been published. In this paper we review these restrictions and prove
some lower-bounds of the computational complexity of learning. In particular,
we show that a learning algorithm for i2-determinate Horn clauses
(with variable i) could be used to decide the PSPACE-complete problem
of Finite State Automata Intersection, and that a learning algorithm for
12-nondeterminate Horn clauses could be used to decide the NP-complete
problem of Boolean Clause Satisfiability (SAT)....
160.
Pharmacophore Discovery using the Inductive Logic Programming System Progol
- Paul Finn,David Page,Ronny Kohavi,Foster Provost
This paper is a case study of a machine aided knowledge discovery process within the
general area of drug design. More specifically, the paper describes a sequence of experiments in
which an Inductive Logic Programming(ILP) system is used for pharmacophore discovery. Within
drug design, a pharmacophore is a description of the substructure of a ligand (a small molecule)
which is responsible for medicinal activity. This medicinal activity is produced by interaction
between the ligand and a binding site on a target protein. ILP was chosen by the domain expert
(first author) at Pfizer since active molecules are most naturally described, in relational terms, as
requiring a substructure...
