Neil Traft

Profile

I am a PhD student in artificial intelligence at the Complex Systems Institute at University of Vermont, advised by Prof. Nick Cheney. I'm currently working on projects that could enable more decentralized and open-ended forms of deep learning and neural architecture search. I'm interested in collective intelligence and self-organizing systems.

Before my PhD, I obtained an MSc in Computer Science from the University of British Columbia, where I studied robotic planning under uncertainty under Ian Mitchell. I've spent a lot of my career working on autonomous vehicles and data science, mostly at Uber ATG (now part of Aurora Innovation). In addition to being a research engineer, I also have some years of experience as a general software engineer as well as a manager of a research engineering team.

Publications

Reset It and Forget It: Relearning Last-Layer Weights Improves Continual and Transfer Learning

Reset It and Forget It: Relearning Last-Layer Weights Improves Continual and Transfer Learning

Lapo Frati, Neil Traft, Jeff Clune, Nick Cheney

European Conference on Artificial Intelligence 2023

ABSHTMLPDF
OmnImage: Evolving 1k Image Cliques for Few-Shot Learning

OmnImage: Evolving 1k Image Cliques for Few-Shot Learning

Lapo Frati, Neil Traft, Nick Cheney

Annual Conference on Genetic and Evolutionary Computation 2023

ABSPDF
Beelines: Motion Prediction Metrics for Self-Driving Safety and Comfort

Beelines: Motion Prediction Metrics for Self-Driving Safety and Comfort

Skanda Shridhar, Yuhang Ma, T. Stentz, Zhengdi Shen, G. C. Haynes, Neil Traft

IEEE International Conference on Robotics and Automation 2021

ABSPDF
Beelines: Evaluating Motion Prediction Impact on Self-Driving Safety and Comfort

Beelines: Evaluating Motion Prediction Impact on Self-Driving Safety and Comfort

Skanda Shridhar, Yuhang Ma, T. Stentz, Zhengdi Shen, G. C. Haynes, Neil Traft

arXiv.org 2020

ABSPDF
Improved action and path synthesis using Gradient Sampling

Improved action and path synthesis using Gradient Sampling

Neil Traft, Ian M. Mitchell

IEEE Conference on Decision and Control 2016

ABS
Educational Robots: Three Models for the Research of Learning Theories and Human-Robot Interaction

Educational Robots: Three Models for the Research of Learning Theories and Human-Robot Interaction

S. Tejada, Neil Traft, M. Hutson, Harold Bufford, M. Dooner, Joshua Hanson, Anthony Radler, George Mauer

AAAI Workshop 2006