ANKUR NATH

Bryan,Texas anath@tamu.edu

I'm Ankur Nath, a motivated PhD student with a strong interest in developing novel Reinforcement Learning (RL) algorithms to solve complex combinatorial optimization problems efficiently. My research focus is to design scalable RL techniques that can provide feasible solutions for larger and more complex optimization problems.

Education

Texas A&M University

Doctor of Philosophy
Computer Science and Engineering

Courses: Artificial Intelligence, Machine Learning, Deep Learning, Deep Reinforcement Learning, Analysis of Algorithms

Current GPA: 4 on a scale of 4

August 2021 - Present

Bangladesh university of engineering and technology

Bachelor of Science
Electrical and Electronic Engineering

CGPA: 3.76 on a scale of 4(top 15%)

February 2015 - April 2019

Research Experience

RL‑Guided Optimization

Focused on developing scalable and interpretable reinforcement learning algorithms to tackle large‑scale combinatorial optimization problems. The research aims to create a versatile framework that maintains high performance across diverse applications while off ering inter‑ pretability. Additionally, the project explores the relationship between a neural network’s complexity and the algorithm it emulates, shedding light on the trade‑off s between performance and interpretability in reinforcement learning.

Noteworthy Projects

SketchyFace

Proposed a deep image-to-image translation method that generates high-quality face images from rough and incomplete freehand sketches

RST-Style Discourse Parsing

Incorporated reinforcement learning in discourse parsing to mitigate error propagation in greedy parsers. Presented an A2C agent that learns non-greedy actions by exploring different state configurations, without relying on gold structure. Conducted experiments with several types of rewards to analyze their effect on the agent's learning.

Sketch2Algo

Development of a visualization tool that instructors can use in their classrooms to teach, explain, and demo machine learning algorithms, and then students can go home and use it to further their understanding

Zlp Scheduler

Developed an application for Zachary Leadership Program (ZLP) to find a time slot for a lecture that fits the schedules of all students within a cohort using an algorithm that finds an open time slot between all schedules

Real-Time Hand Detection and Tracking

Designed a system using real-time hand gestures to control wheelchair mobility, leveraging the advanced Faster R-CNN with minimal image data.

Automated Self Targeting Laser Turret

Design and hardware implementation of a turret that can detect and track intruders

Optical Character Recognition

Distinguish printed text characters inside digital images

Human eye Detection

Detecting the face and eyes of a static image in the frontal plane

Skills

Programming Languages & Tools

Python,C,C++,SQL,Ruby,Assembly Language,MATLAB

Machine Learning framework

Pytorch,Keras,Tensorflow

Dataset visualization tool

Tableau,Seaborn,Matplotlib

Web Development

Node.js, React, HTML, CSS, Dash, Django,Ruby on Rails