Code Repository for IR Research Project as part of TU Delft DSAIT4050 - Group 33

How to recreate results

Noise_Types_Per_Batch.ipynb - Contains all the graphs from our paper. Compares the performance on of different noise levels and types and different batches of Character Error Rate (CER) vs Word Error Rate (WER). All other files were helpful for us to derive the results in this one.
Colbert_Final_evaluation.ipynb - Runs experiments similiar to Noise_Types_Per_Batch.ipynb but with only some subgraphs created.
Document_noise.ipynb - An attempt to assess the feasibility of adding noise to documents and not only queries
model_comparison.ipynb - Compares the performance of various simple Pyterrier retrievers
noise_types.ipynb - A much more detailed analysis of noise types, some of which not included in the final experiment
original_example.ipynb - A sample starting file on how to use frameworks

Name		Name	Last commit message	Last commit date
Latest commit History 29 Commits
.gitignore		.gitignore
Colbert_Final_evaluation.ipynb		Colbert_Final_evaluation.ipynb
Document_noise.ipynb		Document_noise.ipynb
Noise_Types_Per_Batch.ipynb		Noise_Types_Per_Batch.ipynb
README.md		README.md
common_helpers.py		common_helpers.py
model_comparison.ipynb		model_comparison.ipynb
noise_types.ipynb		noise_types.ipynb
original_example.ipynb		original_example.ipynb
plotting_helpers.py		plotting_helpers.py