Akash Sen - Research & Projects

Research Focus

My current research interests lie in Data-Driven learning methods, Inverse Problems and Optimization Techniques.

Publication

Akash Sen, Shantam Gulati, Phanindra Jampana, C.S. Sastry, ``Deep LRST: A learned nonnegative sparse solver with convergence guarantees and application in EIT," submitted to Inverse Problems.
Akash Sen, C.S. Sastry, ``Deep Unrolled Networks for Nonnegative Least Squares Problem: Formulation and Convergence Gurantees," submitted to IEEE Signal Processing Letters.
Akash Sen, C.S. Sastry, ``Learned Over-Parametrized Gradient Descent Method for Non-Negative Least Squares Problem," accepted to SampTA, 2025.
Akash Sen, P. Pradhan, R. Randhi, and C.S. Sastry, ``Learned ReLU-based Soft Thresholding: A data-driven method for non-negative sparse signal recovery," published in IEEE-ICASSP, 2025.
Akash Sen, P. Pradhan, R. Naidu, and C. S. Sastry, ``Unrolled Proximal Gradient Descent Method for Non-Negative Least Squares Problem," published in IEEE-ICASSP, 2024.

Course project for MA 5120 (Numerical Linear Algebra)

Google’s PageRank Algorithm : Bringing Order To The Web

Explored the theoretical & computational aspects of Google’s seminal PageRank algorithm and how numerical linear algebra plays a role in it. Implemented the algorithm in MATLAB. Slides can be found here.

Course project for MA 6040 (Fuzzy Logic Connectives : Theory & Applications)

Fuzzy Relative Position in Image Processing using Morphological Approach

On finding the degree of positional relationship between fuzzy objects - Studied how spatial arrangement of objects in the point of view of fuzzy logic provides important information in contexts like remote sensing, medical imaging etc. Implemented in Python. Slides can be found here.

Master’s project - A study on the analytical aspect of conservation laws

Scalar One Dimensional Conservation Law

Studied different forms of conservation laws and analyzed the behavior of its characteristics. Also came across the Riemann problem and finally investigated some numerical schemes for scalar one dimensional conservation laws such as First-order Godunov Scheme.