Will Cunningham
Head of Quantum Software
Agnostiq
Biography
Will Cunningham leads the software engineering team at Agnostiq in Toronto, Ontario. He designs and implements software for quantum computers and cloud HPC environments. He maintains Covalent as well as the Causal Set Generator and FastMath scientific toolkits. His interests span a wide range, from parallel algorithms and machine learning to quantum gravity and geometry.
Announcements
Covalent is Released!
Covalent is a Pythonic tool for orchestrating HPC and quantum workflows. Covalent helps users match software to the most cost-effective backend hardware, implements checkpointing, manages experimental results, and facilitates reproducible science. Learn more on GitHub and ReadTheDocs.
Will Cunningham leads the software engineering team at Agnostiq in Toronto, Ontario. He designs and implements software for quantum computers and cloud HPC environments. He maintains Covalent as well as the Causal Set Generator and FastMath scientific toolkits. His interests span a wide range, from parallel algorithms and machine learning to quantum gravity and geometry.
Announcements
Covalent is Released!
Covalent is a Pythonic tool for orchestrating HPC and quantum workflows. Covalent helps users match software to the most cost-effective backend hardware, implements checkpointing, manages experimental results, and facilitates reproducible science. Learn more on GitHub and ReadTheDocs.
Vita
Agnostiq (2021 — Present)
Head of Quantum Software
Focus: Distributed Workflows, Cloud Architecture, System Design
Head of Quantum Software
Focus: Distributed Workflows, Cloud Architecture, System Design
Perimeter Institute for Theoretical Physics (2018 — 2021)
Postdoctoral Fellow
Research Focus: Computational Quantum Gravity
Postdoctoral Fellow
Research Focus: Computational Quantum Gravity
Northeastern University (2013 — 2018)
Ph.D. Physics, M.S. Physics
Dissertation: High Performance Algorithms for Quantum Gravity and Cosmology
Ph.D. Physics, M.S. Physics
Dissertation: High Performance Algorithms for Quantum Gravity and Cosmology
Rensselaer Polytechnic Institute (2009 — 2013)
B.S. Physics
Concentrations: Lattice QCD, Astrophysics
B.S. Physics
Concentrations: Lattice QCD, Astrophysics
Code
Covalent
Covalent is a pythonic tool for running high performance/quantum workflows on advanced computing hardware. Learn more on GitHub.
Covalent is a pythonic tool for running high performance/quantum workflows on advanced computing hardware. Learn more on GitHub.
Covalent Plugins
Covalent's plugins allow users to interface Covalent with a variety of backends, including Slurm, Kubernetes, AWS Batch, Amazon Braket, and more. Read more on ReadTheDocs.
Covalent's plugins allow users to interface Covalent with a variety of backends, including Slurm, Kubernetes, AWS Batch, Amazon Braket, and more. Read more on ReadTheDocs.
Beehive
Beehive is a Bash CLI tool designed to create and manage multi-user cloud HPC environments. This project is stable but unreleased; inquire for details.
Beehive is a Bash CLI tool designed to create and manage multi-user cloud HPC environments. This project is stable but unreleased; inquire for details.
Statistical Graph Toolkit
The Statistical Graph Toolkit (Stag) is an SDK used for generating and analyzing random graph ensembles. It includes interfaces in Bash, Python, and C++. This project is in alpha phase and will be released at some point in the near future.
The Statistical Graph Toolkit (Stag) is an SDK used for generating and analyzing random graph ensembles. It includes interfaces in Bash, Python, and C++. This project is in alpha phase and will be released at some point in the near future.
Causal Set Generator
Generate and analyze causal sets and other random geometric graphs using ultra-efficient parallel techniques. Learn more on Bitbucket.
Generate and analyze causal sets and other random geometric graphs using ultra-efficient parallel techniques. Learn more on Bitbucket.
FastMath Toolkit
This toolkit provides optimized mathematical functions and compact data structures along with other useful utilities. Learn more on Bitbucket.
This toolkit provides optimized mathematical functions and compact data structures along with other useful utilities. Learn more on Bitbucket.
Graph Curvature Toolkit
This toolkit provides a set of highly optimized algorithms to calculate the Ollivier-Ricci curvature for graphs. Learn more on Bitbucket.
This toolkit provides a set of highly optimized algorithms to calculate the Ollivier-Ricci curvature for graphs. Learn more on Bitbucket.
Causal Set Classifier
This tutorial uses Covalent and TensorFlow to classify causal sets via supervised learning. Learn more on ReadTheDocs.
This tutorial uses Covalent and TensorFlow to classify causal sets via supervised learning. Learn more on ReadTheDocs.
Tutorials for Optimization
A series of code samples which cover architecture-related optimizations. Download
A series of code samples which cover architecture-related optimizations. Download
Tutorials for Computational Linear Algebra
A series of code samples which cover linear algebra packages in C/C++. Download
A series of code samples which cover linear algebra packages in C/C++. Download
Tutorials for Computational Geometry
A series of code samples which cover methods in computational geometry. Download
A series of code samples which cover methods in computational geometry. Download
Tutorials for Parallel Programming
A series of code samples which cover basic parallel programming techniques. Download
A series of code samples which cover basic parallel programming techniques. Download
HPC Data Manager
A bash tool for managing datasets across multiple HPC clusters. Learn more on Bitbucket.
A bash tool for managing datasets across multiple HPC clusters. Learn more on Bitbucket.
Research
Scalable Graph Algorithms
"Graphs are used to model the relationships between objects or data. Given the widespread availability of CPU and GPU cores in modern computer clusters, we focus on developing novel algorithms which utilize cutting-edge hardware to the fullest extent. Efficient algorithms allows us to model larger systems in shorter times, improve the accuracy of results, predict the arrival of new data, and better identify hidden patterns in experimental data."
Read more here.
"Graphs are used to model the relationships between objects or data. Given the widespread availability of CPU and GPU cores in modern computer clusters, we focus on developing novel algorithms which utilize cutting-edge hardware to the fullest extent. Efficient algorithms allows us to model larger systems in shorter times, improve the accuracy of results, predict the arrival of new data, and better identify hidden patterns in experimental data."
Read more here.
Machine Learning and Artificial Intelligence
"Machine learning and artificial intelligence have become increasingly useful tools for identifying patterns and understanding dynamics in discrete data. We investigate the appliations of these tools to theoretical physics, with a focus on self-assembly of spacetime during the early universe. We also identify connections to models of quantum random walks and quantum Bayesian networks which may be implemented on near-term quantum computers."
Read more here.
"Machine learning and artificial intelligence have become increasingly useful tools for identifying patterns and understanding dynamics in discrete data. We investigate the appliations of these tools to theoretical physics, with a focus on self-assembly of spacetime during the early universe. We also identify connections to models of quantum random walks and quantum Bayesian networks which may be implemented on near-term quantum computers."
Read more here.
Quantum Gravity and Cosmology
"There exist a variety of promising approaches to quantum gravity which involve a discretized model of spacetime. In causal set theory, we study the dynamics of partial orders which represent proto-spacetimes in order to understand how dimension and topology play a role in the gravitational path integral. In the path integral approach to loop quantum gravity, spin foams are coarse-grained in order to identify critical phenomena and continuum limits. And in string theory, we explore the string landscape, represented by a network of geometries, in order to identify selection mechanisms in a multiverse cosmology."
Read more here.
"There exist a variety of promising approaches to quantum gravity which involve a discretized model of spacetime. In causal set theory, we study the dynamics of partial orders which represent proto-spacetimes in order to understand how dimension and topology play a role in the gravitational path integral. In the path integral approach to loop quantum gravity, spin foams are coarse-grained in order to identify critical phenomena and continuum limits. And in string theory, we explore the string landscape, represented by a network of geometries, in order to identify selection mechanisms in a multiverse cosmology."
Read more here.
Discrete Geometry
"The emergence of smooth, continuous space from a limiting sequence of discrete topological objects is called geometrogenesis. We have presented the first proof that there exists a graph curvature which converges to manifold curvature, and further investigated under what conditions this can be measured on the computer. We also measure both qualitatively and quantitatively the boundary geometry of Lorentzian manifolds, with a focus on causally non-convex regions."
Read more here.
"The emergence of smooth, continuous space from a limiting sequence of discrete topological objects is called geometrogenesis. We have presented the first proof that there exists a graph curvature which converges to manifold curvature, and further investigated under what conditions this can be measured on the computer. We also measure both qualitatively and quantitatively the boundary geometry of Lorentzian manifolds, with a focus on causally non-convex regions."
Read more here.
Press
"Software Startup Aims at Cluster Computing Barriers" by Agam Shah, The New Stack (11/2022)
"Agnostiq, which has its roots in high-performance and quantum computing, has upgraded a tool called Covalent as a simple way to partially offload code execution to the best hardware resources in the cloud." Read more here.
"Agnostiq, which has its roots in high-performance and quantum computing, has upgraded a tool called Covalent as a simple way to partially offload code execution to the best hardware resources in the cloud." Read more here.
"New tool for AWS shares workloads across quantum and classical computers" by Ryan Morrison, Tech Monitor (11/2022)
"The middleware tool allows users to access compute resources from both quantum and classical computers regardless of where they are based." Read more here.
"The middleware tool allows users to access compute resources from both quantum and classical computers regardless of where they are based." Read more here.
"Agnostiq Announces Covalent Support for AWS" by Agnostiq, HPCwire (11/2022)
"As more organizations explore the benefits of cloud HPC, tools like Covalent will help them transition without causing disruption to their existing technology stacks." Read more here.
"As more organizations explore the benefits of cloud HPC, tools like Covalent will help them transition without causing disruption to their existing technology stacks." Read more here.
"Start-Up Agnostiq Tackles HPC-Quantum Workflow Development" by John Russell, HPCwire (06/2022)
"Imagine some years hence the HPC landscape, already confusing in its heterogeneity, also includes quantum computing resources. How does a user choose from among those resources the best workflow and computing engines for a project?" Read more here.
"Imagine some years hence the HPC landscape, already confusing in its heterogeneity, also includes quantum computing resources. How does a user choose from among those resources the best workflow and computing engines for a project?" Read more here.
"Navigating the Modern HPC Landscape" by Agnostiq, ISC Tech News (05/2022)
"The HPC landscape is larger, more complex, and more interconnected than ever before. Covalent provides a single interface for users to prototype and scale heterogeneous solutions on any combination of hardware — from a laptop to a supercomputer — and everything in between." Read more here.
"The HPC landscape is larger, more complex, and more interconnected than ever before. Covalent provides a single interface for users to prototype and scale heterogeneous solutions on any combination of hardware — from a laptop to a supercomputer — and everything in between." Read more here.
"Agnostiq Releases Covalent, Open-Source Workflow Orchestration Platform for Quantum and HPC" by Agnostiq, HPCWire (01/2022)
"Covalent is the first open-source workflow orchestration platform designed specifically for quantum and high performance computing." Read more here.
"Covalent is the first open-source workflow orchestration platform designed specifically for quantum and high performance computing." Read more here.
"Agnostiq Selects Pennylane to Develop Quantum Platform for Finance" by Xanadu, PR Newswire (10/2021)
"Agnostiq selects Xanadu's open-source library PennyLane to bring variational quantum computing and quantum machine learning capabilities to financial services." Read more here.
"Agnostiq selects Xanadu's open-source library PennyLane to bring variational quantum computing and quantum machine learning capabilities to financial services." Read more here.
"The Universe Is a Machine That Keeps Learning" by Caroline Delbert, Popular Mechanics (04/2021)
"Basically, we live in one giant algorithm." Read more here.
"Basically, we live in one giant algorithm." Read more here.
"When physicists consult" by Patchen Barss, Perimeter Institute (03/2021)
"The interests of fundamental researchers and tech companies are not so far apart." Read more here.
"The interests of fundamental researchers and tech companies are not so far apart." Read more here.
"Crack the Code" by Stephanie Keating, Perimeter Institute (01/2021)
"Theoretical physics is progressing beyond the blackboard as more researchers develop scientific computing tools and software." Read more here.
"Theoretical physics is progressing beyond the blackboard as more researchers develop scientific computing tools and software." Read more here.
"Northeastern team uses string theory to explain the fundamental nature of the universe" by Laura Castañón, Northeastern University (03/2019)
"Imagine a boiling pot of water, with bubbles perpetually forming and rising to the top. Imagine that within each bubble, there is a universe. And for each of these universes, the laws of physics are slightly different." Read more here. Read featured paper here.
"Imagine a boiling pot of water, with bubbles perpetually forming and rising to the top. Imagine that within each bubble, there is a universe. And for each of these universes, the laws of physics are slightly different." Read more here. Read featured paper here.
Talks
Covalent: Distributed Workflows for Heterogeneous Computing — NYC Quantum Computing Meetup (09/2022)
A technical presentation on distributed workflows in the context of quantum computing.
A technical presentation on distributed workflows in the context of quantum computing.
Panel Talk: Vendor Showdown — HPC & AI on Wall Street (09/2022)
A five-minute lightning talk and panel discussion about emerging HPC technologies.
A five-minute lightning talk and panel discussion about emerging HPC technologies.
Covalent: Distributed Workflows for Heterogeneous Computing — Government of Canada (08/2022)
A semi-technical presentation on distributed workflows in the context of quantum computing. Slides available upon request.
A semi-technical presentation on distributed workflows in the context of quantum computing. Slides available upon request.
Covalent: Distributed Workflows for Heterogeneous Computing — CDL Quantum Bootcamp (07/2022)
A semi-technical presentation on distributed workflows in the context of quantum computing. Slides available upon request.
A semi-technical presentation on distributed workflows in the context of quantum computing. Slides available upon request.
ResearchOps: Challenges and Practical Solutions for Distributed Scientific Computing — Rensselaer Polytechnic Institute (07/2022)
A 90-minute technical presentation on research operations and Covalent, including a hands-on molecular dynamics tutorial. Geared toward a graduate computational science audience. Slides available upon request.
A 90-minute technical presentation on research operations and Covalent, including a hands-on molecular dynamics tutorial. Geared toward a graduate computational science audience. Slides available upon request.
Managing High-Performance Classical and Quantum Workflows with Covalent — Quantum.Tech (06/2022)
A technical lunch-and-learn about Covalent geared toward a senior engineering audience. Slides available upon request.
A technical lunch-and-learn about Covalent geared toward a senior engineering audience. Slides available upon request.
Covalent: Distributed Workflows for Heterogeneous Computing — Quantum.Tech (06/2022)
Overview of Covalent as a product geared toward CTO/CIO advanced technologies leaders. Slides available upon request.
Overview of Covalent as a product geared toward CTO/CIO advanced technologies leaders. Slides available upon request.
High Performance Classical-Quantum Workflows with Covalent — International Supercomputing Conference (05/2022)
We discuss how Covalent can be used to orchestrate software across highly heterogeneous compute frameworks and devices. Video
We discuss how Covalent can be used to orchestrate software across highly heterogeneous compute frameworks and devices. Video
Advances in Quantum Software — Keynote, Inside Quantum Technology (05/2022)
We discuss recent advances in quantum software as well as some of the open challenges this decade. Slides/video available upon request.
We discuss recent advances in quantum software as well as some of the open challenges this decade. Slides/video available upon request.
Why Computer Architecture Matters for HPC — Rensselaer Polytechnic Institute (06/2021)
A survey and tutorial of optimization techniques for C/C++. Slides
A survey and tutorial of optimization techniques for C/C++. Slides
Curvature: From Graphs to Manifolds — Agnostiq (10/2020)
We discuss random geometric graphs, discrete curvature, and recent results in discrete geometry. Slides
We discuss random geometric graphs, discrete curvature, and recent results in discrete geometry. Slides
Computational Geometry for Quantum Gravity — Rensselaer Polytechnic Institute (06/2020)
We review computational methods for discrete quantum gravity. Slides
We review computational methods for discrete quantum gravity. Slides
Quantum Dynamics of Total Orders — CP3-Origins (04/2020)
We discuss quantum growth models in terms of total order dynamics. Slides
We discuss quantum growth models in terms of total order dynamics. Slides
Classical and Quantum Growth Models for Discrete Spacetime — Los Alamos (12/2019)
We discuss models of growth for causal set spacetimes. Slides
We discuss models of growth for causal set spacetimes. Slides
Dimensionally Restricted Causal Sets — Radboud U. (09/2019)
We present evidence of a first order phase transition in dimensionally and topologically restricted causal sets. Slides
We present evidence of a first order phase transition in dimensionally and topologically restricted causal sets. Slides
Timelike Hypersurfaces in Causal Sets — Quantum and Gravity in Okinawa (07/2019)
It is possible to identify and measure timelike boundaries in small causal diamonds. Slides
It is possible to identify and measure timelike boundaries in small causal diamonds. Slides
An Overview of Computational Linear Algebra — Rensselaer Polytechnic Institute (07/2019)
A survey and tutorial of linear algebra packages in C/C++ and Python. Slides
A survey and tutorial of linear algebra packages in C/C++ and Python. Slides
Why Computer Architecture Matters for HPC — Rensselaer Polytechnic Institute (07/2019)
A survey and tutorial of optimization techniques for C/C++. Slides
A survey and tutorial of optimization techniques for C/C++. Slides
Inference of Boundaries in Causal Sets — U. Heidelberg and NetSci2018 (06/2018)
A new algorithm identifies and measures timelike boundaries in (1+1)-dimensional Minkowski causal sets. Slides
A new algorithm identifies and measures timelike boundaries in (1+1)-dimensional Minkowski causal sets. Slides
Deep Learning in Quantum Gravity — Quantum Gravity on the Computer (03/2018)
Deep supervised learning can be used to classify causal sets by manifold and dimension. Slides
Deep supervised learning can be used to classify causal sets by manifold and dimension. Slides
Vacuum Selection from Cosmology Using Networks of String Geometries — Rensselaer Polytechnic Institute (01/2018)
A network model of string geometries shows evidence of vacuum selection in a late-time bubble cosmology. Slides
A network model of string geometries shows evidence of vacuum selection in a late-time bubble cosmology. Slides
The Big Data Approach to Quantum Gravity — Perimeter Institute (12/2017)
Efficient parallel algorithms allow us to better study large problems in causal set theory and F theory. Slides Video
Efficient parallel algorithms allow us to better study large problems in causal set theory and F theory. Slides Video
Introduction to Network Science — String Data Workshop (11/2017)
Network science is useful for modeling many complex systems we observe in the world. Slides
Network science is useful for modeling many complex systems we observe in the world. Slides
Timelike Boundary Terms in the Causal Set Action — Raman Research Institute (12/2016)
Timelike boundaries can be identified in causal sets, and perhaps measured. Slides
Timelike boundaries can be identified in causal sets, and perhaps measured. Slides
Recovering the Einstein-Hilbert Action from Lorentzian Random Geometric Graphs — NetSci2016 (05/2016)
The Benincasa-Dowker action may contain interesting contributions from boundaries. Slides Poster
The Benincasa-Dowker action may contain interesting contributions from boundaries. Slides Poster
An Introduction to Parallel Programming: OpenMP, SSE/AVX, and MPI — Northeastern U. (04/2016)
Parallel computational techniques are easy to incorporate into graph problems. Slides
Parallel computational techniques are easy to incorporate into graph problems. Slides
The Global Airline Network: A Study of Competing Contagions — Northeastern U. (04/2015)
We use a SITR model to study the simultaneous spread of plague, smallpox, and Marburg virus over the airline network. Slides
We use a SITR model to study the simultaneous spread of plague, smallpox, and Marburg virus over the airline network. Slides
Political Influence and Power — Northeastern U. (12/2014)
We study a socio-economic network of politicians constructed from campaign finance data. Slides
We study a socio-economic network of politicians constructed from campaign finance data. Slides
GPU Acceleration for Causal Set Quantum Gravity — Northeastern U. (07/2014)
GPUs are useful tools for generating causal set sprinklings. Slides
GPUs are useful tools for generating causal set sprinklings. Slides