Ashish Kumar


I'm at Tesla working on their humanoid robot (Optimus!).


I graduated from UC Berkeley where I was advised by Prof. Jitendra Malik. Before that, I was a Research Fellow at Microsoft Research India, where I worked with Dr. Manik Varma and Dr. Prateek Jain on developing Resource Efficient Machine Learning algorithms. I am broadly interested in Robotics, with a focus on developing efficient deployable algorithms for locomotion and navigation. I completed my undergraduate study at Indian Institute of Technology Jodhpur, with a major in Computer Science and Engineering. My Bachelor's Thesis was advised by Dr. Gaurav Bhatnagar

Email  /  CV  /  Google Scholar 

profile photo
  Live Demos
       Setup: Quadruped robot climbs on planks, sofa,and stairs then back down.
       Achieved completely from onboard cameras and compute without any preprogrammed gaits!
    At CVPR'22 (New Orleans, USA)
      

    At CoRL'22 (Auckland, New Zealand)
      
  Invited Talks
  • Learning to Walk via Rapid Adaptation: MIT (Computational Sensorimotor Learning Seminar), Stanford (Vision and Learning Lab), UPenn (Grasp Lab Seminar, Kod Lab), UIUC (Computer Vision), University of Washington (RAIVN), UC Berkeley (BAIR Seminar)
  • Visual Memory for Robust Path Following: NeurIPS 2018 (Oral)
  • Learning Navigation Subroutines from Egocentric Videos: UC Berkeley (Semiautonomous Seminar)
  • The Edge of Machine Learning: Facebook AI Research, Oxford University, MSR Cambridge, MSR Redmond, ICML 2017 (short version of the full talk)

  Publications

Legged Locomotion in Challenging Terrains using Egocentric Vision
Ananye Agarwal*, Ashish Kumar*, Jitendra Maliki+, Deepak Pathak+
CoRL (Best Systems Paper Award)

webpage | abstract | bibtex | arXiv

@inproceedings{agarwal2022legged,
  title={Legged locomotion in 
  challenging terrains 
  using egocentric vision},
  author={Agarwal, Ananye and Kumar, Ashish 
  and Malik, Jitendra and Pathak, Deepak},
  booktitle={6th Annual Conference 
  on Robot Learning}
}

Learning Visual Locomotion with Cross-Modal Supervision
Antonio Loquercio*, Ashish Kumar*, Jitendra Malik
arXiv Preprint

webpage | abstract | bibtex | arXiv

@article{loquercio2022learning,
  title={Learning Visual Locomotion 
  with Cross-Modal Supervision},
  author={Loquercio, Antonio and Kumar, Ashish 
  and Malik, Jitendra},
  booktitle={arXiv preprint arXiv:2211.07638}
  year={2022}
}

In-Hand Object Rotation via Rapid Motor Adaptation
Haozhi Qi*, Ashish Kumar*, Roberto Calandra, Yi Ma, Jitendra Malik
CoRL 2022

webpage | abstract | bibtex | arXiv

@inproceedings{qihand,
  title={In-Hand Object Rotation 
  via Rapid Motor Adaptation},
  author={Qi, Haozhi and Kumar, Ashish 
  and Calandra, Roberto and 
  Ma, Yi and Malik, Jitendra},
  booktitle={6th Annual Conference 
  on Robot Learning}
}

A Zero-shot Adaptive Quadcopter Controller
Dingqi Zhang, Antonio Loquercio, Xiangyu Wu,
Ashish Kumar, Jitendra Malik, Mark W. Mueller
arXiv Preprint

webpage | abstract | bibtex | arXiv

@article{zhang2022zero,
  title={A Zero-Shot Adaptive 
  Quadcopter Controller},
  author={Zhang, Dingqi and Loquercio, Antonio 
  and Wu, Xiangyu and Kumar, Ashish and 
  Malik, Jitendra and Mueller, Mark W},
  journal={arXiv preprint arXiv:2209.09232},
  year={2022}
}

Adapting Rapid Motor Adaptation for Bipedal Robots
Ashish Kumar*, Zhongyu Li*, Jun Zeng, Deepak Pathak,
Koushil Sreenath, Jitendra Malik
IROS 2022

webpage | abstract | bibtex | arXiv | demo

@article{arma,
  title={Adapting Rapid Motor
  Adaptation for Bipedal Robots},
  author={Kumar, Ashish and Li,
  Zhongyu and Zeng, Jun and Pathak,
  Deepak and Sreenath, Koushil
  and Malik, Jitendra},
  journal={IROS},
  year={2022}
}

Coupling Vision and Proprioception for Navigation of Legged Robots
Zipeng Fu*, Ashish Kumar*, Ananye Agarwal, Haozhi Qi,
Jitendra Malik, Deepak Pathak
CVPR 2022
Best Paper Award in Multimodal Learning Workshop

Minimizing Energy Consumption Leads to the Emergence of Gaits in Legged Robots
Zipeng Fu, Ashish Kumar, Jitendra Malik, Deepak Pathak
CoRL 2021

webpage | pdf | abstract | bibtex

@article{fu2021minimizing,
  author = {Fu, Zipeng and
  Kumar, Ashish and Malik, Jitendra
  and Pathak, Deepak},
  title  = {Minimizing Energy
  Consumption Leads to the Emergence
  of Gaits in Legged Robots},
  journal= {Conference on Robot Learning (CoRL)},
  year   = {2021}
}

RMA: Rapid Motor Adaptation for Legged Robots
Ashish Kumar, Zipeng Fu, Deepak Pathak, Jitendra Malik
RSS 2021

webpage | pdf | abstract | bibtex | arXiv |

@article{kumar2021rma,
  author = {Kumar, Ashish and
  Fu, Zipeng and Pathak, Deepak
  and Malik, Jitendra},
  title  = {RMA: Rapid Motor
  Adaptation for Legged Robots},
  journal= {RSS},
  year   = {2021}
}

Learning Navigation Subroutines from Egocentric Videos
Ashish Kumar, Saurabh Gupta, Jitendra Malik
CoRL 2019

webpage | pdf | abstract | bibtex | arXiv

@article{kumar2019learning,
  author = {Kumar, Ashish and 
  Gupta, Saurabh and Malik, Jitendra},
  title  = {Learning navigation 
  subroutines from egocentric videos},
  journal= {Conference on Robot Learning (CoRL)},
  year   = {2019}
}

Visual Memory for Robust Path Following
Ashish Kumar*, Saurabh Gupta*, David Fouhey, Sergey Levine, Jitendra Malik
NeurIPS 2018

webpage | pdf | abstract | bibtex | arXiv

@article{kumar2018visual,
  author = {Kumar, Ashish 
  and Gupta, Saurabh and Fouhey, David 
  and Levine, Sergey and Malik, Jitendra},
  title  = {Visual Memory for 
  Robust Path Following},
  journal= {NeurIPS},
  year   = {2018}
}
sym

FastGRNN: A Fast, Accurate, Stable and Tiny Kilobyte Sized Gated Recurrent Neural Network
Aditya Kusupati, Manish Singh, Kush Bhatia, Ashish Kumar, Prateek Jain, Manik Varma
NeurIPS 2018

Repository | pdf | abstract | bibtex | arXiv

@article{kusupati2018fastgrnn,
  author = {Kusupati, Aditya and 
  Singh, Manish and Bhatia, Kush and 
  Kumar, Ashish and Jain, Prateek and 
  Varma, Manik},
  title  = {Fastgrnn: A fast, accurate, stable 
  and tiny kilobyte sized gated 
  recurrent neural network},
  journal= {NeurIPS},
  year   = {2018}
}
sym

Resource-efficient Machine Learning in 2 KB RAM for the Internet of Things
Ashish Kumar, Saurabh Goyal, Manik Varma
ICML 2017

Repository | pdf | abstract | bibtex

@article{kumar2017resource,
  author = {Kumar, Ashish 
  and Goyal, Saurabh and Varma, Manik},
  title  = {Resource-efficient machine 
  learning in 2 KB RAM for 
  the internet of things},
  journal= {ICML},
  year   = {2017}
}
sym

ProtoNN: Compressed and Accurate kNN for Resource-scarce Devices
Chirag Gupta, Arun Sai Suggala, Ankit Goyal, Harsha Vardhan Simhadri, Bhargavi Paranjape, Ashish Kumar, Saurabh Goyal, Raghavendra Udupa, Manik Varma, Prateek Jain
ICML 2017

Repository | pdf | abstract | bibtex

@article{gupta2017resource,
  author = {Gupta, Chirag 
  and Suggala, Arun Sai and 
  Goyal, Ankit and Simhadri, Harsha Vardhan and 
  Paranjape, Bhargavi and Kumar, Ashish and 
  Goyal, Saurabh and Udupa, Raghavendra and 
  Varma, Manik and Jain, Prateek},
  title  = {Protonn: Compressed and accurate 
  knn for resource-scarce devices},
  journal= {ICML},
  year   = {2017}
}


  Teaching Experience
  • GSI for CS280: Computer Vision -- Spring 2019
  • GSI for CS194-26: Image Manipulation, Vision and Comp. Photo -- Spring 2020
  • Education Support Volunteer Make a Difference Foundation -- Sept 2016 - May 2017

Personal

Scuba Diving : Great Barrier Reef, Australia (10m); Phuket, Thailand (12m)
Sky dive : Lompoc, California -- 13000 feet
Bungy Jump : Rishikesh, India -- 83m


Website adapted from Jon Barron, Deepak Pathak and Saurabh Gupta