David Millard

GitHub / LinkedIn / CV / Google Scholar

I am currently building a robotics company. If this interests you in any way, email me at my first name @azalearobotics.com.

I am a roboticist and software engineer with a background in mathematics and computer science. I previously worked at Google X on the Everyday Robots project and at Microsoft.

I completed a Ph.D. in 2023 at the University of Southern California, advised by Gaurav Sukhatme. My research focused on the use of differentiable physics and learning for robotic systems manipulating non-rigid objects. I was funded by a NASA Space Technology Research Fellowship and worked extensively with the Intelligent Robotics Group at Ames Research Center, as well as the Jet Propulsion Lab. During my PhD, I also interned at Google DeepMind and at Iron Ox.

My goal is to create Robots of Economic Value that Don’t Make the World Worse (REVDMWWs).

Research interests

• Robotic systems in dynamic environments
• Fast physics simulation for motion planning and optimal control
• Strong inductive biases in machine learning

Other interests

• C++ and Julia
• Bicycles
• Radios (KN6SAW)

Conference Publications

• David Millard, Daniel Pastor, Joseph Bowkett, Paul Backes and Gaurav S Sukhatme. “GranularGym: High Performance Simulation for Robotic Tasks with Granular Materials,” presented at the Robotics: Science and Systems (Daegu, Korea). 2023. ( arXiv, pdf, show BibTeX )

  @inproceedings{millard2023granulargym,
    title={GranularGym: High Performance Simulation for Robotic Tasks with Granular Materials},
    author={Millard, David and Pastor, Daniel and Bowkett, Joseph and Backes, Paul and Sukhatme, Gaurav S},
    eventtitle={Robotics: Science and Systems},
    venue={Daegu, Korea},
    year={2023},
    month=jul
  }

• David Millard, James A Preiss, Jernej Barbič and Gaurav S Sukhatme. “Parameter Estimation for Deformable Objects in Robotic Manipulation Tasks,” presented at the International Symposium on Robotics Research (Geneva, Switzerland). 2022. ( show BibTeX )

  @inproceedings{millard2022parameter,
    title={Parameter Estimation for Deformable Objects in Robotic Manipulation Tasks},
    author={Millard, David and Preiss, James A and Barbič, Jernej and Sukhatme, Gaurav S},
    eventtitle={International Symposium on Robotics Research},
    venue={Geneva, Switzerland},
    year={2022},
    month=sep
  }

• Eric Heiden, Christopher E Denniston, David Millard, Fabio Ramos and Gaurav S Sukhatme. “Probabilistic Inference of Simulation Parameters via Parallel Differentiable Simulation,” presented at the IEEE Conference on Robotics and Automation (Philadelphia, USA). 2022. ( IEEE, arXiv, pdf, show BibTeX )

  @inproceedings{heiden2021probabilistic,
    title={Probabilistic Inference of Simulation Parameters via Parallel Differentiable Simulation},
    author={Heiden, Eric and Denniston, Christopher E and Millard, David and Ramos, Fabio and Sukhatme, Gaurav S},
    eventtitle={IEEE Conference on Robotics and Automation},
    venue={Philadelphia, USA},
    year={2022},
    month=may
  }

• James A Preiss*, David Millard*, Tao Yao* and Gaurav S Sukhatme. “Tracking Fast Trajectories with a Deformable Object using a Learned Model,” presented at the IEEE Conference on Robotics and Automation (Philadelphia, USA). 2022. ( IEEE, pdf, show BibTeX )

  @inproceedings{preiss2022tracking,
    title={Tracking Fast Trajectories with a Deformable Object using a Learned Model},
    author={Preiss, James A and Millard, David and Yao, Tao and Sukhatme, Gaurav S},
    eventtitle={IEEE Conference on Robotics and Automation},
    venue={Philadelphia, USA},
    year={2022},
    month=may
  }

• Eric Heiden*, David Millard*, Erwin Coumans, Yizhou Sheng and Gaurav S Sukhatme. “NeuralSim: Augmenting differentiable simulators with neural networks,” presented at the IEEE International Conference on Robotics and Automation (Xi'an, China). 2021. ( IEEE, arXiv, pdf, show BibTeX )

  @inproceedings{heiden2021neuralsim,
    title={NeuralSim: Augmenting differentiable simulators with neural networks},
    author={Heiden, Eric and Millard, David and Coumans, Erwin and Sheng, Yizhou and Sukhatme, Gaurav S},
    eventtitle={IEEE International Conference on Robotics and Automation},
    venue={Xi'an, China},
    year={2021},
    month=sep
  }

Workshop Papers

• David Millard, James A Preiss, Jernej Barbič and Gaurav S Sukhatme. “Direct System Identification of Deformable Objects using Differentiable Finite Element Dynamics,” presented at the IROS 2022 3rd Workshop on Robotic Manipulation of Deformable Objects (ROMADO-SI) (Kyoto, Japan). 2022. ( show BibTeX )

  @inproceedings{millard2022direct,
    title={Direct System Identification of Deformable Objects using Differentiable Finite Element Dynamics},
    author={Millard, David and Preiss, James A and Barbič, Jernej and Sukhatme, Gaurav S},
    eventtitle={IROS 2022 3rd Workshop on Robotic Manipulation of Deformable Objects (ROMADO-SI)},
    venue={Kyoto, Japan},
    year={2022},
    month=oct
  }

• James A Preiss*, David Millard*, Tao Yao* and Gaurav S Sukhatme. “Deep Recurrent Models for Nonlinear Model Predictive Control in Deformable Manipulation Tasks,” presented at the ICRA 2022 2nd Workshop on Representing and Manipulating Deformable Objects (Philadelphia, USA). 2022. ( show BibTeX )

  @inproceedings{preiss2022deep,
    title={Deep Recurrent Models for Nonlinear Model Predictive Control in Deformable Manipulation Tasks},
    author={Preiss, James A and Millard, David and Yao, Tao and Sukhatme, Gaurav S},
    eventtitle={ICRA 2022 2nd Workshop on Representing and Manipulating Deformable Objects},
    venue={Philadelphia, USA},
    year={2022},
    month=may
  }

• Eric Heiden*, David Millard*, Erwin Coumans and Gaurav S Sukhatme. “Sparse-Input Neural Network Augmentations for Differentiable Simulators,” presented at the NeurIPS 2020 Workshop on Differentiable Computer Vision, Graphics, and Physics in Machine Learning (Virtual). 2020. ( show BibTeX )

  @inproceedings{heiden2020sparse,
    title={Sparse-Input Neural Network Augmentations for Differentiable Simulators},
    author={Heiden, Eric and Millard, David and Coumans, Erwin and Sukhatme, Gaurav S},
    eventtitle={NeurIPS 2020 Workshop on Differentiable Computer Vision, Graphics, and Physics in Machine Learning},
    venue={Virtual},
    year={2020},
    month=dec
  }

• Eric Heiden, David Millard and Gaurav Sukhatme. “Real2Sim Transfer using Differentiable Physics,” presented at the RSS 2019 Workshop on Closing the Reality Gap in Sim2real Transfer for Robotic Manipulation (Freiburg, Germany). 2019. ( show BibTeX )

  @inproceedings{heiden2019real2sim,
    title={Real2Sim Transfer using Differentiable Physics},
    author={Heiden, Eric and Millard, David and Sukhatme, Gaurav},
    eventtitle={RSS 2019 Workshop on Closing the Reality Gap in Sim2real Transfer for Robotic Manipulation},
    venue={Freiburg, Germany},
    year={2019},
    month=jun
  }

Preprints

• Eric Heiden*, David Millard*, Erwin Coumans and Gaurav S Sukhatme. “Augmenting Differentiable Simulators with Neural Networks to Close the Sim2Ueal Gap,” arXiv preprint arXiv:2007.06045. 2020. ( arXiv, pdf, show BibTeX )

  @article{heiden2020augmenting,
    title={Augmenting Differentiable Simulators with Neural Networks to Close the Sim2Ueal Gap},
    author={Heiden, Eric and Millard, David and Coumans, Erwin and Sukhatme, Gaurav S},
    journal={arXiv preprint arXiv:2007.06045},
    year={2020},
    month=jul
  }

• David Millard, Eric Heiden, Shubham Agrawal and Gaurav S Sukhatme. “Automatic Differentiation and Continuous Sensitivity Analysis of Rigid Body Dynamics,” arXiv preprint arXiv:2001.08539. 2020. ( arXiv, pdf, show BibTeX )

  @article{millard2020automatic,
    title={Automatic Differentiation and Continuous Sensitivity Analysis of Rigid Body Dynamics},
    author={Millard, David and Heiden, Eric and Agrawal, Shubham and Sukhatme, Gaurav S},
    journal={arXiv preprint arXiv:2001.08539},
    year={2020},
    month=jan
  }

• Eric Heiden*, David Millard*, Hejia Zhang and Gaurav S Sukhatme. “Interactive Differentiable Simulation,” arXiv preprint arXiv:1905.10706. 2019. ( arXiv, pdf, show BibTeX )

  @article{heiden2019interactive,
    title={Interactive Differentiable Simulation},
    author={Heiden, Eric and Millard, David and Zhang, Hejia and Sukhatme, Gaurav S},
    journal={arXiv preprint arXiv:1905.10706},
    year={2019},
    month=may
  }

Condensed history

I am now building a robotics company. If this interests you in any way, email me at my first name @azalearobotics.com.

In 2022, I was a visiting technologist at the Caltech/NASA Jet Propulsion Lab in group 347G, working on autonomous science for robotic probes sampling the surfaces of moons of the outer planets.

In 2020, I was a visiting technologist at NASA Ames Research Center, working with the Astrobee team in the Intelligent Robotics Group.

In 2020, I was a research intern at Google Brain, working on differentiable physics and learning for robots with high inductive bias.

In 2019, I worked with Iron Ox on kinodynamically limited arm planning problems using trajectory optimization methods.

From 2015 to 2018, I was a software engineer at Google X working on robotics.

From 2014 to 2015, I was a software engineer at Microsoft.

I graduated Summa Cum Laude in 2014 with a B.S. in mathematics and computer science from the University of Georgia, where I studied under the auspices of the Foundation Fellowship.

In July 2013, I worked in Khayelitsha, South Africa, collaborating with the Schools Improvement Initiative at the University of Cape Town to repair and improve computer labs at primary and secondary schools through free software and open knowledge bases, like Wikipedia.

From May to August 2012, I researched multiagent mapping systems in the Computer Networks and Telematics Group at the University of Freiburg, in Freiburg, Germany.

In 2012 and 2013, I served as president of UGA Mathcounts Outreach.

From 2009 to 2013, I was a staff member of Free IT Athens.