J
Jörg Lücke
Researcher at University of Oldenburg
Publications - 77
Citations - 1125
Jörg Lücke is an academic researcher from University of Oldenburg. The author has contributed to research in topics: Neural coding & Artificial neural network. The author has an hindex of 17, co-authored 75 publications receiving 997 citations. Previous affiliations of Jörg Lücke include Technical University of Berlin & Ruhr University Bochum.
Papers
More filters
Journal ArticleDOI
Beyond Manual Tuning of Hyperparameters
TL;DR: This work discusses two strategies towards making machine learning algorithms more autonomous: automated optimization of hyperparameters (including mechanisms for feature selection, preprocessing, model selection, etc) and the development of algorithms with reduced sets ofhyperparameters.
Journal ArticleDOI
Hysteresis effects on the input–output curve of motor evoked potentials
TL;DR: Findings implicate that hysteresis can influence IO curves significantly, and one possible underlying mechanism might be short-term synaptic enhancement.
Journal Article
Maximal Causes for Non-linear Component Extraction
Jörg Lücke,Maneesh Sahani +1 more
TL;DR: It is shown that learning in recent softmax-like neural networks may be interpreted as approximate maximization of a data likelihood, and results of learning model parameters to fit acoustic and visual data sets in which max-like component combinations arise naturally.
Journal ArticleDOI
Rapid processing and unsupervised learning in a model of the cortical macrocolumn
TL;DR: Minicolumns are shown to be able to organize their collective inputs without supervision by Hebbian plasticity into selective receptive field shapes, thereby becoming classifiers for input patterns and demonstrating its ability for distributed neural coding.
Journal Article
Expectation Truncation and the Benefits of Preselection In Training Generative Models
Jörg Lücke,Julian Eggert +1 more
TL;DR: It is shown how a preselection of hidden variables can be used to efficiently train generative models with binary hidden variables and found that the training scheme can reduce computational costs by orders of magnitude and allows for a reliable extraction of hidden causes.