The work was done as part of my Undergraduate Final Year Project and is greatly influenced by "Wasserstein Distributionally Robust Learning. Shafieezadeh Abadeh, SorooshKuhn, Daniel (Dir.), EPFL, Lausanne, 2020".
During this project we worked on understanding the formulations for DRO in context of common machine learning algorithms of Support Vector Machine and Logistic Regression.
The work closely followed the following works :
- "Wasserstein Distributionally Robust Learning. Shafieezadeh Abadeh, SorooshKuhn, Daniel (Dir.), EPFL, Lausanne, 2020"
- Mohajerin Esfahani, P., Kuhn, D. Data-driven distributionally robust optimization using the Wasserstein metric: performance guarantees and tractable reformulations. Math. Program. 171, 115–166 (2018). https://doi.org/10.1007/s10107-017-1172-1
We worked on implementing the formulations mentioned in (2).
The work was done in with Siddharth Singh Solanki as part of the degree requirement for BTech from IIT Goa, under the guidance of Dr. Divya Padmanabhan