These are notebooks associated with Kaggle competitions some of our memebers have participated in. The models are adaptations of well known arcitectures and are used to solve the problems presented in the competitions. The notebooks are listed below. Click the button to access the notebooks.
This is an encoder-decoder style transformer model used to detect and translate American Sign Language (ASL) fingerspelling into text: asl.ipynb
This notebook uses pretrained bird vocalization classifier from Google to classify bird sounds: birdClef.ipynb
This notebook uses BERT based classifier to evaluate essays writtes by Eglish language learners: ell.ipynb
This notebook uses a Temporal Fusion Transformer to predict monthly microbusiness density in a given area: tft.ipynb
This notebook uses a Temporal Fusion Transformer to forecast return rates of investments: ubq.ipynb