An Eecient Validation Mechanism for Inductive Logic Programming Using Compositionality
نویسنده
چکیده
Inductive Logic Programming, which consists in learning clauses from examples, can be viewed as a cycle conception/validation leading to the acceptance of the induced program provided that it fulllls a certain criterion. We focus on the validation step in the context of empirical multi-predicate learning of normal clauses. Thanks to a compositional semantics, the classical validation step of the complete induced program can be replaced by the veriication of local properties for a cut out into units, considerably limiting the usual combinatorial explosion. Moreover, we provide a semantics-preservative transformation which allows to simplify the program and provides a further reenement of the cut out. La Programmation Logique Inductive consiste a apprendre des clauses a partir d'exemples et peut ^ etre vue comme un cycle conception/validation menant a l'acceptation du programme induit d es qu'il satisfait un certain crit ere. Nous nous int eressons plus particuli erement a l' etape de validation dans le contexte de l'apprentissage empirique multi-pr edicat de clauses normales. Gr^ ace a une s emantique compositionelle, l' etape classique de la validation du programme com-plet peut ^ etre remplac ee par la v eriication de propri et es locales pour un certain d ecoupage en unit es, ce qui limite consid erablement l'explosion combinatoire. De plus, nous proposons une transformation pr eservant la s emantique qui permet de simpliier le programme et propose un raanement du d ecoupage.
منابع مشابه
Verification of Logic Programs and Imperative Programs
This paper explores the relationship between veriication of logic programs and imperative programs with the aim of uncovering the kinds of reasoning used to construct logic programs. We discuss forward reasoning, such as that used for verifying imperative programs using the inductive assertion method, and backward reasoning, such as that used for verifying imperative programs using subgoal indu...
متن کاملInductive Veri cation and Validation of Multi-Agent Systems
Most veriication and validation (V&V) methods employ deductive reasoning to verify whether an implementation is in agreement with a speciication. We present a novel approach to veriication and validation of complex systems that is based on inductive reasoning. Induction allows to derive general rules from speciic behaviours of the software (e.g. the inputs and outputs). Using inductive logic pr...
متن کاملUsing ILP-Systems for Verification and Validation of Multi-agent Systems
Most applications of inductive logic programming focus on prediction or the discovery of new knowledge. We describe a less common application of ILP namely veriication and validation of knowledge based systems and multi-agent systems. Using inductive logic programming, partial declarative speciications of the software can be induced from the behavior of the software. These rules can be readily ...
متن کاملCompositional Relational Programming with Name Projection and Compositional Synthesis
CombInduce is a methodology for inductive synthesis of logic programs, which employs a reversible meta-interpreter for synthesis, and uses a compositional relational target language for efficient synthesis of recursive predicates. The target language, Combilog, has reduced usability due to the lack of variables, a feature enforced by the principle of compositionality, which is at the core of th...
متن کاملMultiple-Instance Learning Based Heuristics for Mining Chemical Compound Structure
Inductive Logic Programming (ILP) is a combination of inductive learning and first-order logic aiming to learn first-order hypotheses from training examples. ILP has a serious bottleneck in an intractably enormous hypothesis search space. This makes existing approaches perform poorly on large-scale real-world datasets. In this research, we propose a technique to make the system handle an enormo...
متن کامل