ScaDS Logo

COMPETENCE CENTER
FOR SCALABLE DATA SERVICES
AND SOLUTIONS

TSO-1: Deep Learning

 

Service Owner

ZIH, Method scientists

Contact

Sunna Torge, Peter Winkler

Target

Data analysts

Dependencies

None.

Description

Deep learning is a promising and fast developing field in machine learning allowing the hierarchical composition of arbitrary mathematical functions to map any input onto output data. This general formulation enables various use cases ranging from the prediction of future time series data to the recognition of latent patterns in audio and image data. The accuracy of various network architectures strongly depends on the input data, let it convolutional networks showing best performance for images, recurrent nets/LSTMs for time series and sensor data or deep autoencoders for efficient data compression. We support the whole chain of designing a Deep Learning system tailored to the presented use case up to deploy prototypic solutions that ensure optimal performance.

Offerings

  •  Provide hardware infrastructure for distributed Deep Learning methods.

  •  Provide frameworks for various distributed Deep Learning networks.

  •  Support the selection of an optimal architecture for given use case.

  •  Support the design and implementation of necessary (pre)processing methods to deploy a Deep Learning architecture.

  •  Support the qualitative and quantitative evaluation of the chosen design.

Consumption

  •  Collect information about your use case. Prepare for the following questions:

    •  Who is responsible for your project?

    •  What task should be solved?

    •  How would you describe your data?

    •  Do you already have a (serial/parallel) program?

    •  What time and resource constraints do you have?

  •  Contact us (via e-mail or phone)

  •  We send you an application form. With this form, we want to have a look at your use case and see the specific requirements. This helps us to provide any additional software you might need. Additionally, we need this form to request computing resources.

  •  Fill out the form and send it back to us.

  •  We contact you and discuss your use case.