[Comp-neuro] OrGanic Environment for Reservoir Computing (OGER) toolbox v1.0 released

David Verstraeten david.verstraeten at elis.ugent.be
Sun Oct 10 15:31:28 CEST 2010


We are glad to announce the first official release of OGER (OrGanic Environment for Reservoir computing), a Python toolbox for rapidly building, training and evaluating modular learning architectures on large datasets. 

OGER is available from http://organic.elis.ugent.be/oger .

OGER builds functionality on top of the well-known Modular toolkit for Data Processing (MDP). Through MDP, Oger provides:
• Easily building, training and using modular structures of learning algorithms
• A wide variety of state-of-the-art machine learning methods, such as PCA, ICA, SFA, RBMs, ..

The Oger toolbox adds functionality to this such as:
• Several additional nodes such as Reservoir Computing nodes, a ridge regression node, a Conditional RBM node and a perceptron node
• Easy parallelization on computing clusters
• GPU-based acceleration 
• Cross-validation and grid-searching of large parameter spaces
• Processing of sequential or temporal datasets
• Recursive generation of sequences
• Gradient-based training of (deep) learning architectures
• Interface to the Speech Processing, Recognition, and Automatic Annotation Kit (SPRAAK) 
• Interface to PyNN-compatible spiking neural network simulators

Several tutorials and example datasets are provided to demonstrate the capabilities of OGER. The tutorials include:
- Modeling of place-cells for a robot with a reservoir with SFA and ICA
- Classification of MNIST data using a Deep-Belief Network with perceptron readout
- Constructing and training a TRM (Temporal Reservoir Machine) modular architecture
- ...

The OGER toolbox has been funded by the EU FP7 project ORGANIC.


----------------------------------------
Dr. ir. David Verstraeten
Reservoir lab
Department of Electronics and Information Systems
Ghent University

Website:
http://snn.elis.ugent.be/david

Address:
Sint Pietersnieuwstraat 41
B-9000 Ghent, Belgium
Phone : +32 9 264 34 04
Fax: + 32 9 264 35 94








-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.neuroinf.org/pipermail/comp-neuro/attachments/20101010/c53ff582/attachment.html


More information about the Comp-neuro mailing list