Before You Read "Graph of Thoughts"
A comprehensive primer on understanding Graph of Thoughts (GoT) paper, covering what came before GoT, the mathematical intuition, and how to approach graph-based reasoning in LLMs.
(he/him)
Ph.D. Student in Biomedical Informatics & Data Science
I am Naif A. Ganadily, a Ph.D. student in the Biomedical Informatics and Data Science program at Arizona State University (ASU). I am working as a graduate research associate and AI Research Intern at Mayo Clinic under the supervision of Dr. Irbaz Riaz (ASU/Mayo Clinic) and Prof. Li Liu.
Before starting my PhD at ASU, I completed my Master’s degree in Electrical Engineering (Machine Learning & Data Science) at the University of Washington in 2024, where I worked with Prof. Stan Birchfield and collaborated with NVIDIA on computer vision and synthetic data generation research. Prior to that, I completed my undergraduate degree in Electrical Engineering from the University of Business and Technology in Jeddah, Saudi Arabia in 2021.
I am interested in Explainable AI, Large Language Models (LLMs), Generative AI applications in healthcare, AI-driven clinical decision support systems, Privacy-Preserving Machine Learning, and Computer Vision. My current research focuses on developing LLM and rule-based systems for hematology oncology and creating pipelines to automatically identify Major Adverse Cardiac Events (MACE) in clinical notes.
At Mayo Clinic, I’m working on:
I also write about AI and machine learning on Medium, where I’ve published articles explaining complex AI concepts like Graph of Thoughts prompting for large language models.
I am fluent in English and Arabic, and hold dual citizenship in the USA and Saudi Arabia.
Ph.D. in Biomedical Informatics and Data Science
Arizona State University
M.S. in Electrical Engineering (Machine Learning & Data Science)
University of Washington
B.S. in Electrical Engineering
University of Business and Technology
A comprehensive primer on understanding Graph of Thoughts (GoT) paper, covering what came before GoT, the mathematical intuition, and how to approach graph-based reasoning in LLMs.
Orchestrating an end-to-end Generative AI and Rule Based system pipeline to replicate the full triage system of Classical Hematology Oncology
An end-to-end Generative AI system pipeline to detect MACE outcomes in Clinical Notes
Implementing differential privacy and pseudonymization in TensorFlow Federated to protect patient data while maintaining model performance
Developed synthetic datasets using Blender/nvisii to improve LoFTR model accuracy by 12% and robustness by 10%