Tyler Streeter

Theoretical Machine Learning / Statistics Researcher: Bayes, Information Theory, Boltzmann Machines

Iowa City, Iowa, United States of America

Research Interests

artificial intelligence

machine learning

undirected graphical models

Boltzmann machines

Markov random fields

Ising models

spin glasses

Hopfield networks

energy-based models

log-linear models

directed graphical models

feedforward neural networks

Bayesian networks

logistic belief networks

self-organizing maps

generative models

kernel mixture models

optimization algorithms

gradient descent

coordinate descent

2nd-order optimization methods

natural gradient

simulated annealing

genetic and evolutionary algorithms

deep learning

MCMC sampling algorithms

Gibbs sampling

Metropolis-Hastings

exact/perfect sampling

partition function/free energy estimation

unsupervised learning

reinforcement learning

clustering

lossless data compression

text/sequence prediction

information theory

Bayesian methods

objective Bayesian priors

maximum entropy

minimum description length (MDL)

Fisher information

computational neuroscience

brain-inspired AI

artificial curiosity

About

I am a researcher and software engineer focused on making machine learning simpler, more general, and more effective. Having spent many years studying a wide range of existing models and algorithms, I now work on deriving new methods from elegant theoretical principles. I enjoy writing clean code and simple APIs, designing data visualizations to gain intuition about new domains, simulating physical processes with unexpected emergent behavior, building tangible objects from humble materials, and capturing big ideas with small math. My ideal project is one that lets me be a scientist, artist, and engineer.

Publications

Verve: A General Purpose Open Source Reinforcement Learning Toolkit

Volume 1: 32nd Design Automation Conference, Parts A and B / Jan 01, 2006

Streeter, T., Oliver, J., & Sannier, A. (2006, January 1). Verve: A General Purpose Open Source Reinforcement Learning Toolkit. Volume 1: 32nd Design Automation Conference, Parts A and B. https://doi.org/10.1115/detc2006-99651

Design and implementation of general purpose reinforcement learning agents

Streeter, T. E. (n.d.). Design and implementation of general purpose reinforcement learning agents [Iowa State University]. https://doi.org/10.31274/rtd-20201107-491

Education

Iowa State University

PhD (ABD), Machine Learning / May, 2024 (anticipated)

Ames, Iowa, United States of America

Iowa State University

MS, Reinforcement Learning, Computer Graphics / December, 2005

Ames, Iowa, United States of America

Iowa State University

BS, Computer Engineering / May, 2004

Ames, Iowa, United States of America

Experience

Brainpower Labs

Machine Learning Researcher / October, 2008 — Present

• Pure AI/ML research and software development. • AI research/development contract with SRAM. • Derived math results (currently 6,500 pages of notes), and designed new learning algorithms involving probabilistic graphical models, Bayesian methods, and information theory. • Built internal software tools in C/C++ and Python to aid research, including interactive visualizations of machine learning and Monte Carlo sampling algorithms. • Designed a novel brain-inspired architecture for artificial general intelligence, and implemented it in in C++ and Python with interactive debugger and test environments. • Developed commercial software to fund research agenda, including iBonsai, a meditative interactive 3D tree simulation in C++ for iOS (120k users). • Graphics engineering contract with Avatree (custom generative 3D tree growth algorithm and glTF exporter in C).

VR Applications Center, Iowa State University

AI/ML Graduate Researcher / August, 2006 — December, 2009

• Performed independent research on topographic maps, maximum entropy learning algorithms, Bayesian networks, reinforcement learning, and systems neuroscience. • Developed open source C++ libraries for unit testing, profiling, and parallel programming.

IBM Research

Computational Neuroscience Research Intern / May, 2006 — August, 2006

• Implemented a novel computational model of the cerebellum. • Demonstrated motor learning and transfer of complex reaching behaviors with a simulated 6-muscle arm. • Participated in discussions of global brain modeling and information theoretic learning rules.

Iowa State University

Teaching Assistant / January, 2006 — May, 2006

• Assisted teaching machine perception techniques, computer vision, image processing, and software tools (e.g. OpenCV). • Managed wiki for student code submissions.

VR Applications Center, Iowa State University

AI/ML Graduate Researcher / May, 2003 — December, 2005

• Performed independent research for MS thesis on reinforcement learning, artificial curiosity, planning, 3D physics simulation, and interactive real-time visualization. • Developed virtual reality software tools and applications for several funded research projects.

Center for Nondestructive Evaluation, Iowa State University

3D Graphics Research Assistant / August, 2002 — December, 2003

• Assisted design and implementation of 3D software for ultrasonic probe simulations used for testing e.g. jet engine and space shuttle parts. • Integrated STL file importer and octree space partitioning.

Ames National Laboratory

HPC Research Assistant / June, 2002 — August, 2002

• Developed general message logging software for computational chemistry and other applications on high performance computing clusters.

Links & Social Media

tyl.st (personal website)

Join Tyler on NotedSource!

Join Now

At NotedSource, we believe that professors, post-docs, scientists and other researchers have deep, untapped knowledge and expertise that can be leveraged to drive innovation within companies. NotedSource is committed to bridging the gap between academia and industry by providing a platform for collaboration with industry and networking with other researchers.

For industry, NotedSource identifies the right academic experts in 24 hours to help organizations build and grow. With a platform of thousands of knowledgeable PhDs, scientists, and industry experts, NotedSource makes connecting and collaborating easy.

For academic researchers such as professors, post-docs, and Ph.D.s, NotedSource provides tools to discover and connect to your colleagues with messaging and news feeds, in addition to the opportunity to be paid for your collaboration with vetted partners.