the proceedings contain 20 papers. the topics discussed include: building theories of the world: human and machine learning perspectives;SRL without tears: an ilp perspective;semantic web meets ilp: unconsumated love,...
ISBN:
(纸本)3540859276
the proceedings contain 20 papers. the topics discussed include: building theories of the world: human and machine learning perspectives;SRL without tears: an ilp perspective;semantic web meets ilp: unconsumated love, or no love lost?;learning expressive models of gene regulation;information overload and FP7 funding opportunities in 2009-10;a model to study phase transition and plateaus in relational learning;top-down induction of relational model trees in multi-instance learning;challenges in relational learning for real-time systems applications;discriminative structure learning of Markov logic networks;an experiment in robot discovery withilp;using the bottom clause and mode declarations on FOL theory revision from examples;DL-FOIL: concept learning in description logics;feature discovery with type extension trees;and feature construction using theory-guided sampling and randomised search.
the proceedings contain 28 papers. the topics discussed include: learning with kernels and logical representations;beyond prediction: directions for probabilistic and relational learning;learning probabilistic logic m...
详细信息
ISBN:
(纸本)3540784683
the proceedings contain 28 papers. the topics discussed include: learning with kernels and logical representations;beyond prediction: directions for probabilistic and relational learning;learning probabilistic logic models from probabilistic examples;learning directed probabilistic logical models using ordering-search;learning to assign degrees of belief in relational domains;bias/variance analysis for relational domains;induction of optimal semantic semi-distances for clausal knowledge bases;clustering relational data based on randomized propositionalization;structural statistical software testing with active learning in a graph;learning declarative bias;learning relational options for inductive transfer in relational reinforcement learning;a phase transition-based perspective on multiple instance kernels;and combining clauses with various precisions and recalls to produce accurate probabilistic estimates.
We revisit an application developed originally using abductive inductivelogicprogramming (ilp) for modeling inhibition in metabolic networks. the example data was derived from studies of the effects of toxins on rat...
详细信息
We revisit an application developed originally using abductive inductivelogicprogramming (ilp) for modeling inhibition in metabolic networks. the example data was derived from studies of the effects of toxins on rats using Nuclear Magnetic Resonance (NMR) time-trace analysis of their biofluids together with background knowledge representing a subset of the Kyoto Encyclopedia of Genes and Genomes (KEGG). We now apply two Probabilistic ilp (Pilp) approaches-abductive Stochastic logic Programs (SLPs) and programming In Statistical modeling (PRISM) to the application. Both approaches support abductive learning and probability predictions. Abductive SLPs are a Pilp framework that provides possible worlds semantics to SLPs through abduction. Instead of learning logic models from non-probabilistic examples as done in ilp, the Pilp approach applied in this paper is based on a general technique for introducing probability labels within a standard scientific experimental setting involving control and treated data. Our results demonstrate that the Pilp approach provides a way of learning probabilistic logic models from probabilistic examples, and the Pilp models learned from probabilistic examples lead to a significant decrease in error accompanied by improved insight from the learned results compared withthe Pilp models learned from non-probabilistic examples.
the field of inductivelogicprogramming (ilp) has made steady progress, since the first ilp workshop in 1991, based on a balance of developments in theory, implementations and applications. More recently there has be...
详细信息
the field of inductivelogicprogramming (ilp) has made steady progress, since the first ilp workshop in 1991, based on a balance of developments in theory, implementations and applications. More recently there has been an increased emphasis on Probabilistic ilp and the related fields of Statistical Relational Learning (SRL) and Structured Prediction. the goal of the current paper is to consider these emerging trends and chart out the strategic directions and open problems for the broader area of structured machine learning for the next 10 years.
the problem of determining the Worse Case Execution Time (WCET) of a piece of code is a fundamental one in the Real Time Systems community. Existing methods either try to gain this information by analysis of the progr...
详细信息
ISBN:
(纸本)9783540859277
the problem of determining the Worse Case Execution Time (WCET) of a piece of code is a fundamental one in the Real Time Systems community. Existing methods either try to gain this information by analysis of the program code or by running extensive timing analyses. this paper presents a new approach to the problem based on using Machine Learning in the form of ilp to infer program properties based on sample executions of the code. Additionally, significant improvements in the range of functions learnable and the time taken for learning can be made by the application of more advanced ilp techniques.
thrs special issue is associated withthe 16thinternationalconference of inductivelogicprogramming (ilp 2006), which represented a radical departure from previous years. Submissions were requested in two phases. T...
详细信息
thrs special issue is associated withthe 16thinternationalconference of inductivelogicprogramming (ilp 2006), which represented a radical departure from previous years. Submissions were requested in two phases. the first phase involved submission of short papers (3 pages) which were then presented at the conference and included in a short papers proceedings. In the second phase, reviewers selected papers for long paper submission (15 pages maximum). these were then assessed by the same reviewers, who then decided which papers to include in the Journal Special Issue and Proceedings. In the first phase there were a record 77 papers, compared to the usual 20 or so long papers of previous years. Each paper was reviewed by 3 reviewers. Out of these, 71 were invited to submit long papers. Out of the long paper submissions, 7 were selected for the Machine Learning Journal Special Issue and 27 were accepted for the conference proceedings. In addition, two papers were nominated by PC referees for the applications prize and two for the theory prize. the papers in this special issue represent the diversity and vitality in present ilp research.
Stochastically searching the space of candidate clauses is an appealing way to scale up ilp to large datasets. We address an approach that uses a Bayesian network model to adaptively guide search in this space. We exa...
详细信息
ISBN:
(纸本)9783540784685
Stochastically searching the space of candidate clauses is an appealing way to scale up ilp to large datasets. We address an approach that uses a Bayesian network model to adaptively guide search in this space. We examine guiding search towards areas that previously performed well and towards areas that ilp has not yet thoroughly explored. We show improvement in area under the curve for recall-precision curves using these modifications.
We identify a shortcoming of a standard positive-only clause evaluation function within the context of learning biological grammars. To overcome this shortcoming we propose L-modification, a modification to this evalu...
详细信息
ISBN:
(纸本)9783540859277
We identify a shortcoming of a standard positive-only clause evaluation function within the context of learning biological grammars. To overcome this shortcoming we propose L-modification, a modification to this evaluation function such that the lengths of individual examples are considered. We use a set of bio-sequences known as neuropeptide precursor middles (NPP-middles). Using L-modification to learn from these NPP-middles results in induced grammars that have a better performance than that achieved when using the standard positive-only clause evaluation function. We also show that L-modification improves the performance of induced grammars when learning on short, medium or long NPPs-middles. A potential disadvantage of L-modification is discussed. Finally, we show that, as the limit on the search space size increases, the greater is the increase in predictive performance arising from L-modification.
暂无评论