This volume contains the text of the five invited papers and16 selected contributions presented at the thirdInternational Workshop on Analogical and InductiveInference, AII `92, held in Dagstuhl Castle, Germany,October 5-9, 1992.Like the two previous events, AII '92 was intended to bringtogether representatives from several research communities,in particular, from theoretical computer science, artificialintelligence, and from cognitive sciences.The papers contained in this volume constitute astate-of-the-art report on formal approaches to algorithmiclearning, particularly emphasizing aspects of analogicalreasoning and inductive inference. Both these areas arecurrently attracting strong interest: analogical reasoningplays a crucial role in the booming field of case-basedreasoning, and, in the fieldof inductive logic programming,there have recently been developed a number of newtechniques for inductive inference.
Representing the spatial/kinematic domain and lattice computers.- A solution of the credit assignment problem in the case of learning rectangles.- Learning decision strategies with genetic algorithms.- Background knowledge and declarative bias in inductive concept learning.- Too much information can be too much for learning efficiently.- Some experiments with a learning procedure.- Unions of identifiable classes of total recursive functions.- Learning from multiple sources of inaccurate data.- Strong separation of learning classes.- Desiderata for generalization-to-N algorithms.- The power of probabilism in Popperian FINite learning.- An analysis of various forms of `jumping to conclusions´.- An inductive inference approach to classification.- Asking questions versus verifiability.- Predictive analogy and cognition.- Learning a class of regular expressions via restricted subset queries.- A unifying approach to monotonic language learning on informant.- Characterization of finite identification.- A model of the `redeillegalscription´ process in the context of geometric proportional analogy problems.- Inductive strengthening: The effects of a simple heuristic for restricting hypothesis space search.- On identifying DNA splicing systems from examples.
This volume contains the text of the five invited papers and16 selected contributions presented at the thirdInternational Workshop on Analogical and InductiveInference, AII `92, held in Dagstuhl Castle, Germany,October 5-9, 1992.Like the two previous events, AII '92 was intended to bringtogether representatives from several research communities,in particular, from theoretical computer science, artificialintelligence, and from cognitive sciences.The papers contained in this volume constitute astate-of-the-art report on formal approaches to algorithmiclearning, particularly emphasizing aspects of analogicalreasoning and inductive inference. Both these areas arecurrently attracting strong interest: analogical reasoningplays a crucial role in the booming field of case-basedreasoning, and, in the fieldof inductive logic programming,there have recently been developed a number of newtechniques for inductive inference.
Representing the spatial/kinematic domain and lattice computers.- A solution of the credit assignment problem in the case of learning rectangles.- Learning decision strategies with genetic algorithms.- Background knowledge and declarative bias in inductive concept learning.- Too much information can be too much for learning efficiently.- Some experiments with a learning procedure.- Unions of identifiable classes of total recursive functions.- Learning from multiple sources of inaccurate data.- Strong separation of learning classes.- Desiderata for generalization-to-N algorithms.- The power of probabilism in Popperian FINite learning.- An analysis of various forms of 'jumping to conclusions'.- An inductive inference approach to classification.- Asking questions versus verifiability.- Predictive analogy and cognition.- Learning a class of regular expressions via restricted subset queries.- A unifying approach to monotonic language learning on informant.- Characterization of finite identification.- A model of the 'redeillegalscription' process in the context of geometric proportional analogy problems.- Inductive strengthening: The effects of a simple heuristic for restricting hypothesis space search.- On identifying DNA splicing systems from examples.
Inhaltsverzeichnis
Representing the spatial/kinematic domain and lattice computers.- A solution of the credit assignment problem in the case of learning rectangles.- Learning decision strategies with genetic algorithms.- Background knowledge and declarative bias in inductive concept learning.- Too much information can be too much for learning efficiently.- Some experiments with a learning procedure.- Unions of identifiable classes of total recursive functions.- Learning from multiple sources of inaccurate data.- Strong separation of learning classes.- Desiderata for generalization-to-N algorithms.- The power of probabilism in Popperian FINite learning.- An analysis of various forms of 'jumping to conclusions'.- An inductive inference approach to classification.- Asking questions versus verifiability.- Predictive analogy and cognition.- Learning a class of regular expressions via restricted subset queries.- A unifying approach to monotonic language learning on informant.- Characterization of finite identification.- A model of the 'redeillegalscription' process in the context of geometric proportional analogy problems.- Inductive strengthening: The effects of a simple heuristic for restricting hypothesis space search.- On identifying DNA splicing systems from examples.
Klappentext
This volume contains the text of the five invited papers and
16 selected contributions presented at the third
International Workshop on Analogical and Inductive
Inference, AII `92, held in Dagstuhl Castle, Germany,
October 5-9, 1992.
Like the two previous events, AII '92 was intended to bring
together representatives from several research communities,
in particular, from theoretical computer science, artificial
intelligence, and from cognitive sciences.
The papers contained in this volume constitute a
state-of-the-art report on formal approaches to algorithmic
learning, particularly emphasizing aspects of analogical
reasoning and inductive inference. Both these areas are
currently attracting strong interest: analogical reasoning
plays a crucial role in the booming field of case-based
reasoning, and, in the fieldof inductive logic programming,
there have recently been developed a number of new
techniques for inductive inference.
Springer Book Archives
