Home

Team 09 / Model Distillation

Organization

• Weekly meeting every Monday 11 am. Zoom-Link

Workflow

• Master branch is protected
• Merge to master by create pull request and approval by team members
• SCRUM like process. Use 'Projects' tab
  • Create Issue as a ticket
  • ONLY work on this issue, if another problem arises, open another issue.
  • Create branch from issue
  • If work is fulfilled, create pull request
• Codestyle: Type Hints.
• Docstyle: As described here.

Papers

ML-Papers

• Model Distillation: A Survey
• Positive-Unlabeled Learning with Non-Negative Risk Estimator
• Positive-Unlabeled Learning from Imbalanced Data

Background Knowledge

• Joint epitope selection and spacer design for string-of-beads vaccines
• Integrated modeling of the major events in the MHC class I antigen processing pathway
• The role of the proteasome in generating cytotoxic T-cell epitopes: insights obtained from improved predictions of proteasomal cleavage
• Bounding and Counting Linear Regions of Deep Neural Networks (not part of this project)
• Learning from Corrupted Binary Labels via Class-Probability Estimation
• Recovering True Classifier Performance in Positive-Unlabeled Learning

Useful Links

• Summary of Model Distillation Survey

Code Inspiration

• RCNN Distillation
• Distillation in Keras
• Knowledge Distillation of DNNs
• Promising Implementation
• Training convolutional neural networks with cheap convolutions and online distillation