Open-source implementations and reproducible computational artifacts accompanying selected research papers.
The repositories below provide experiment code, numerical pipelines, configuration files, examples, and saved artifacts for work on stochastic generative models, dependent time series, and Schrödinger–Bass transport.
Schrödinger-Bridge Drift Estimation Experiments
Reproducible experiments for direct nonparametric estimation of Schrödinger-bridge drifts from dependent time-series data. The repository accompanies the finite-sample, asymptotic, and adaptive analysis developed in the associated paper.
- Kernel drift estimators and deterministic reference computations
- Rate, central-limit, adaptivity, and stress-test drivers
- Configurations, processed results, figures, and tests
SBBTS
Official implementation of the Schrödinger–Bass framework for synthetic financial time-series generation and data augmentation.
- End-to-end Heston simulation workflow
- S&P 500 cluster augmentation experiments
- Training, sampling, and preprocessing utilities
LightSBB-M
Official implementation of a lightweight Schrödinger–Bass bridge method for generative transport, including synthetic experiments and an image-translation example.
- Model architecture and training implementation
- Reusable Schrödinger–Bass package
- Example Jupyter notebook for image translation