Dr. Sambhav Jain

I am a researcher in statistical machine learning and optimisation with expertise in multi-label learning, Twin Support Vector Machines, and kernel methods. My research focuses on developing scalable and theoretically grounded optimisation frameworks for large-scale and weakly supervised learning. More recently, I have been working on interpretable machine learning, federated learning, and multi-omics data integration for biomarker discovery.

Research Intern, SciWhy Lab - Jawaharlal Nehru University - School of Computational and Integrative Sciences

(2026-03)

• Conducted research on multi-omics data integration for lung cancer prediction.
• Developed and evaluated interpretable and statistical machine learning models for accurate cancer classification.
• Identification of potential diagnostic biomarkers from high-dimensional multi-omics datasets using feature selection techniques.

Ph.D. Research Scholar, MLSI Lab - South Asian University - Department of Computer Science and Engineering

(2020-11 - 2026-06)

• Proposed a Unified Twin-Loss optimisation framework for Multi-label learning, enabling simultaneous handling of missing labels, noise, and outliers within a single formulation.
• Developed convergence analysis and optimisation guarantees for convex and non-convex loss settings.
• Designed scalable variants using low-rank formulations, DCA and stochastic optimisation techniques (SGD, SVRG), significantly reducing computational and memory costs.
• Extended Twin SVM frameworks to indefinite kernels via polar decomposition, broadening applicability beyond positive semi-definite kernel assumptions.
• Adapted the framework to limited-supervision scenarios, including semi-supervised learning and multi-label learning with missing labels.
• Resulted in publications in IEEE, Springer and Elsevier journals.

Ph.D. - Computer Science and Engineering - South Asian University (2020 - 2026)

M.Sc. - Computer Science - South Asian University (2018 - 2020)

B.Sc. - Computer Science - University of Delhi (2015 - 2018)