This book is the final report on a comprehensive basic research project, named GOSLER on algorithmic learning for knowledge-based systems supported by the German Federal Ministry of Research and Technology during the years 1991 - 1994. This research effort was focused on the study of fundamental learnability problems integrating theoretical research with the development of tools and experimental investigation.
The contributions by 11 participants in the GOSLER project is complemented by contributions from 23 researchers from abroad. Thus the volume provides a competent introduction to algorithmic learning theory.

Inhalt
Learning and consistency.- Error detecting in inductive inference.- Learning from good examples.- Towards reduction arguments for FINite learning.- Not-so-nearly-minimal-size program inference (preliminary report).- Optimization problem in inductive inference.- On identification by teams and probabilistic machines.- Topological considerations in composing teams of learning machines.- Probabilistic versus deterministic memory limited learning.- Classification using information.- Classifying recursive predicates and languages.- A guided tour across the boundaries of learning recursive languages.- Pattern inference.- Inductive learning of recurrence-term languages from positive data.- Learning formal languages based on control sets.- Learning in case-based classification algorithms.- Optimal strategies Learning from examples Boolean equations.- Feature construction during tree learning.- On lower bounds for the depth of threshold circuits with weights from {?1,0,+1}.- Structuring neural networks and PAC-Learning.- Inductive synthesis of rewrite programs.- TLPS A term rewriting laboratory (not only) for experiments in automatic program synthesis.- GoslerP A logic programming tool for inductive inference.