Machine learning applied to conformal field theory
A computer-science-minor capstone project: build neural networks that learn to classify types of quantum field theories from synthetic data, reproducing and extending results from a 2020 physics paper. Co-authored with a fellow student in autumn 2020, as an end-to-end Python pipeline running from synthetic data generation through to trained models.
This was my CS-minor capstone project at NISER Bhubaneswar, done alongside my physics coursework, not a research-lab project.
The problem
Reproduce and extend the machine-learning approach to classifying conformal field theories introduced by Chen, He, Lal and Zaz (2020, arXiv:2006.16114), to see how well the method held up and where it could be pushed further.
The approach
An end-to-end Python pipeline: NumPy for synthetic conformal-field-theory data generation, Keras/TensorFlow multilayer perceptrons (MLPs) as the models, and scikit-learn for data splitting. The project covered seven classification tasks and one regression task in total.
The outcome
A public repository with separate main, data-gen, and model branches, and a 28-slide presentation delivered in December 2020.
The dossier does not give quantified accuracy or error figures for the seven classification tasks or the regression task, only that the pipeline reproduced and extended the reference paper's approach. No numeric results are stated here to avoid inventing figures the source material does not contain.