CellSignal

Chair for Computer Aided Medical Procedures & Augmented Reality
Lehrstuhl für Informatikanwendungen in der Medizin & Augmented Reality

THIS WEBPAGE IS DEPRECATED - please visit our new website

Tasks

Participation in challenges is the funniest and most efficient way to master your AI skill.
We invite candidates (e.g. IDP, guided research students as well as random people) to participate in CellSignal challenge organized in conjunction with NeurIPS2019 conference.
Time framework: 1st June - 1st Nov, 2019

g880.png
"The task is to correctly classify populations of human cells as exhibiting one of 1,108 different gene knockdown conditions. The dataset consists of 125,568 high-resolution 6-channel fluorescence microscopy images of cells under these gene knockdown conditions across 4 cell types and 51 different repeats of the experiment, resulting in images which exhibit technical nuisance factors per experiment called batch effects that confound the classification task. The data is split into training and test sets by experiment repeat, so good classifiers will need to separate relevant biological factors from the batch effect factors in order to generalize outside of the training data."

Requirements

  • Sound knowledge of machine learning and computer vision.
  • Experience in working with deep learning frameworks (PyTorch?, Tensorflow).
  • Competitive spirit.

Contact

Ivan Ezhov
Suprosanna Shit



Edit | Attach | Refresh | Diffs | More | Revision r1.8 - 15 May 2019 - 18:36 - IvanEzhov