As part of finding duplicates (using facial recognition) in a database, I explore an original algorithm which is able to make sure all graphs of duplicates are complete graphs.

Explore how can unstructured documents be turned into structured data suitable for NLP tasks, allowing information communication through extraction.

Explore the results of this practical optimization experiment aiming to curate a challenging dataset for finding the best stack for facial recognition from detection to feature extraction and distance metric.

Explore an ML algorithm which is able to infer text classes from semantic clusters.