MSc in Computing Science
I am a Research Assistant at the University of Alberta working with Martha White in the RLAI Lab.
My current research centers on online reinforcement learning, specifically on creating methods that dynamically tune hyperparameters during agent-environment interactions. I have also contributed to designing continuing environments that evaluate an agent’s adaptability in dynamicaly changing partially observable settings. Additionally, drawing inspiration from animal learning theories, I explore concepts such as environmental enrichment - studying how diverse, task-agnostic observations can enhance an agent’s abilities across both simple and complex RL environments. I am also interested in exploring capacity-constrained, lifelong learning agents, as I believe intelligent behviors arise from mitigating natural limitations, while adapting, planning and interacting with the environment.
Contact: hakhverd [at] ualberta [dot] ca
- Continual Reinforcement Learning
- Single-lifetime Agents
- Capacity-constrained Learning
- Animal-inspired Learning
MSc in Computing Science
University of Alberta
BSc in Information Technologies
National Polytechnic University of Armenia
Experience
Research Assistant
University of Alberta - RLAI LabResponsibilities include:
- Finishing up my thesis work for journal submission.
- Developing a suite of environments for process control.
- Working on developing environments suitable for Continual learning.
- Understanding and evaluating RL agents in curated animal learning experiments.
Machine Learning Engineer
Magical LabsResponsibilities include:
- Developed models and pipelines for autonomous vehicle control.
- Worked on compressing large neural networks with RL agents for mobile integration.
Machine Learning Engineer
Perpetuum MotorsResponsibilities include:
- Conducted research on developing an end-to-end autonomous driving pipeline for simulated games.
Software Engineer
ForgeFictionResponsibilities include:
- Developed and designed API requests and database architectures.
- Optimized the application for maximum speed and scalability.
Education
MSc in Computing Science
University of AlbertaGPA: 3.8/4.0
My Master’s thesis explores online hyperparameter tuning in reinforcement learning, where an agent adapts its hyperparameters during interactions rather than relying on a pre-tuning phase.
BSc in Information Technologies
National Polytechnic University of ArmeniaGPA: 3.92/4.0