Skip to content

Latest commit

 

History

History
22 lines (13 loc) · 1.63 KB

README.md

File metadata and controls

22 lines (13 loc) · 1.63 KB

Master Thesis in AI

If you have any questions regarding my results or work I'm happy to answer them! Just contact me over GitHub or open an issue.

Thesis

For my written master thesis including detailed results, experiment descriptions and theoretical background please refer to my work "Attentive Tabular Learning in Context of Drug Discovery".

Project

This project contains all the source for my master thesis including:

  • tabnet: a pure pytorch based tabnet reimplementation including various attentive types including sparsemax, entmax-15, alpha-entmax and others.
  • tabnet_lightning: a pytorch lightning wrapper for various tasks, at the moment mostly focusing on classification tasks
  • tests: pytest based test suite for all relevant implementations. The TabNet reimplementation is quite rigorously tested.
  • baseline: reference and baseline model code including a MLP, RF and GBDT for comparison used during experiments.
  • datasets: pytorch lighting based data modules for various datasets mostly within the drug discovery field. Focuses on local caching and uses multi-processing wherever possible. Includes various molecule featurizer implementations which also can calculate the atomic contribution of feature dimension (basically the reverse path feature to atom in molecule).

Results

Access to all experiments is provided at mlflow.kriechbaumer.at (might be taken down in the future).