In addition to the many free online tools that SciServer provides for understanding big scientific data, we also provide data and tools for a number of other Partner Applications. This page describes the Partner Applications that connect with SciServer.

List of Partner Applications

This section lists each of SciServer’s current Partner Applications. Click on a name to expand the tab and learn more about each partner application.

DeepForge is a web-based gateway to deep learning for the scientific community. It enables users to visually define neural network architectures along with training or preprocessing workflows, and then execute these workflows on platforms like SciServer Compute. Utilizing built-in version control, multi-user collaboration, and a combination of visual and textual interfaces, DeepForgeis designed to both facilitate the rapid development of neural network models and promote reproducibility of the experiments.

SciServer provides required infrastructure for DeepForge. SciServer Compute provides infrastructure for executing workflows from within DeepForge and the storage of data artifacts such as training data is provided by the Files view of the SciServer Dashboard.

Flowcharts illustrating Deepforge operations
Left: An example training pipeline defined in DeepForge for training a classification model on CIFAR-10.
Right: Monitoring the execution of the pipeline on SciServer from within the browser.
The Tomographer (tomographer.org) is a tool for exploring the third dimension (or redshift distribution) of any source catalog or sky map, using the clustering-redshift technique.

It allows one to explore virtually any dataset: at any wavelength and over a wide range of angular resolutions. The larger the footprint, the better. It is designed to be as simple as possible for the user: simply upload a source catalog or a Healpix intensity map and the Tomographer will infer its redshift distribution.

SciServer provides the infrastructure for launching computational jobs that generate the tomography, as well as the storage space needed for the application and tomography results.

Left: Input HEALPix map
Right: Redshift Tomography