This project was created as part of the Master AI Lab at Karlsruhe University of Applied Sciences. It deals with the Contradictory, My Dear Watson NLP kaggle challenge. We used several approaches such as a probabilistic approach, forest based methods and deep learning methods and compared their results to each other.
Developers:
| Name | ID |
|---|---|
| Alexander Niedermayer | nial1016 |
| Dominik Schneider | scdo1042 |
| Christian Reichhardt | rech1028 |
Our data analysis can be found in this notebook.
A detailed description of our results and next steps can be found here. You can train the best performing model using this command:
python train_final_model.py