- Description
- My Master's thesis, "A Transformer-based Approach to the Book Likeability Prediction Problem", is a research project that harnesses the power of the neural Transformer architecture to assess the quality of books in an automated fashion. Through use of the novel Transformer architecture we were able to yield the best end-to-end performance with a weighted F1 score of 73.6% on the benchmark Maharjan Goodreads dataset which is competitive with the overall State-of-the-Art.
- Tools used
- Huggingface, Ray-tune, Pytorch, WandB
Six Page Extended Abstract
