litlas
ログインアカウント作成

Discovering Predictable Classifications

著者: Jürgen Schmidhuber, Daniel Prelinger - Neural Computation 1993 被引用: 50

Prediction problems are among the most common learning problems for neural networks (e.g., in the context of time series prediction, control, etc.). With many such problems, however, perfect prediction is inherently impossible. For such cases we present novel unsupervised systems that learn to classify patterns such that the classifications are predictable while still being as specific as possible. The approach can be related to the IMAX method of Becker and Hinton (1989) and Zemel and Hinton (1991). Experiments include a binary stereo task proposed by Becker and Hinton, which can be solved more readily by our system.

A mathematical theory ofcommunicationA mathematical theory of communicationSelf-Organization in aPerceptual NetworkSelf-Organization in a Perceptual NetworkFinding Minimum EntropyCodesFinding Minimum Entropy CodesDiscoveringViewpoint-Invariant…Discovering Viewpoint-Invariant Relationships That Characterize ObjectsLearning Factorial Codesby Predictability…Learning Factorial Codes by Predictability MinimizationSubsymbolic naturallanguage processing : A…Subsymbolic natural language processing : An integrated model of scripts, lexicon, and memory. By Risto Miikkulainen. Cambridge: MIT Press, 1993. Pp. xv, 403.Learning Factorial Codesby Predictability…Learning Factorial Codes by Predictability MinimizationThe discovery ofstructure by…The discovery of structure by multi-stream networks of local processors with contextual guidanceActivation Functions,Computational Goals and…Activation Functions, Computational Goals and Learning Rules for Local Processors with Contextual GuidanceIn search of commonfoundations for cortica…In search of common foundations for cortical computationContextually guidedunsupervised learning…Contextually guided unsupervised learning using local multivariate binary processorsLocal online learning ofcoherent informationLocal online learning of coherent informationCategory LearningThrough Multi-Modality…Category Learning Through Multi-Modality SensingFeature ExtractionThrough LOCOCODEFeature Extraction Through LOCOCODEArtificial curiositybased on discovering…Artificial curiosity based on discovering novel algorithmic predictability through coevolutionUnsupervised Learning inLSTM Recurrent Neural…Unsupervised Learning in LSTM Recurrent Neural NetworksExploring thePredictableExploring the PredictableDeep Learning in NeuralNetworks: An OverviewDeep Learning in Neural Networks: An OverviewDiscovering PredictableClassificationsDiscovering Predictable Classifications過去の参考文献中心の論文この論文を引用する論文古い新しい

ノードをクリックするとフォーカスを固定、空白をクリックすると本論文に戻ります。ホバーで一時的にプレビューできます。各ノードのページはタイトルから開けます。