Synthetic Data Generation With A Proof Assistant For Training Large Language Models
Master’s thesis
A meta-programming framework in Lean for synthetic data generation through forward reasoning.
Projects
Molecule Retrieval with Natural Language Queries
Academic project
Data challenge mixing natural language processing and graph neural networks.
CPOP: Continuous-piecewise-linear Pruned Optimal Partitioning
Academic project
Implementation of a changepoint detection method.
Stock Classification from High Frequency Market Data
Academic project
Data challenge on market data with a distribution shift.
Variational EM algorithm for the Stochastic Block Model
Academic project
An implementation of an EM algorithm for the Stochastic Block Model.
Lymphocytosis classification
Academic project
Data challenge on medical image classification.