Student Researcher, DIVE LAB

Actively engaged as a member of the DIVE Lab at Texas A&M University, specializing in cutting-edge research at the intersection of Large Language Models (LLMs), Graph Neural Networks (GNN), and Graph Transformers (GT).

Currently, my scholarly pursuits center around unraveling the intricate internal mechanisms of Graph Models. My research aims to contribute a comprehensive understanding of how Graph Transformers and Large Language Models interpret and process graph datasets.

As part of my master's thesis, I am dedicated to advancing the field and aspire to present my findings at esteemed conferences such as NeurIPS, ICML, and ICLR. My ultimate goal is to contribute valuable insights to the academic community and propel the understanding of these innovative technologies to new heights.

As an active participant in DIVE Lab, I regularly contribute to insightful discussions on compelling research papers during our collaborative sessions, sharing insights and fostering intellectual discussions within our lab community. Explore some of my recent presentations below.

Emergent Abilities of Large Language Models

• Do LLMs really possess Emergent Abilities or are these abilities just a mirage?

TALK LIKE A GRAPH: Encoding Graphs for Large Language Models

• How to make LLMs understand Graph datasets? Study the effect of Graph encoding and structure on LLM Performance

Tree of Thoughts: Deliberate Problem Solving with Large Language Models

• Tree of Thoughts (ToT) generalizes over the popular Chain of Thought prompting and enables exploration over coherent units of text(thoughts) that serve as intermediate steps toward problem solving.

LONGLORA: Efficient fine-tuning of long context Large Language Models

• LongLoRA is an efficient fine-tuning approach that extends the context sizes of pre-trained large language models (LLMs) to 8192 tokens, with limited computation cost.

Recent Blog posts

I avidly keep abreast of the latest developments in the NLP community through regular reading of blog posts on platforms like Medium. Actively engaging with the community, I exchange thoughts and ideas. Occasionally, I contribute my own insights by writing on topics that I find relevant but underexplored. Explore some of my thought-provoking blog posts below.

Explainability in Graph Neural networks with GNNExplainer
NEW

• Why does a trained GNN model predict a certain class label and how to make it predict a desired class label? GNNExplainer is a groundbreaking approach providing interpretable explanations for any GNN-based model.

One For ALL : One Graph Model For All Classification Tasks
NEW

• Developing a unified model for diverse tasks in artificial intelligence was historically challenging for graph datasets. One For All (OFA) is a recent groundbreaking framework that employs a single graph model, showcasing robust performance across various tasks and domains.

Competitive programming

As a passionate coding enthusiast, my journey began with competitive programming during my undergraduate days. Currently, I am an active member on LeetCode, consistently maintaining a global rank within the top 2000 out of 465,000 users. Demonstrating dedication, my current streak spans over 365 days.

I contribute not only through participation but also by providing detailed solutions, particularly for the most challenging problems. Explore a selection of my top-rated articles below, offering a rich source of my coding insights and problem-solving strategies.

Number of Good Paths

• 150+ Likes received• 5000+ views

Collect coins in a tree

• 150+ Likes received• 4000+ views