Ganesh Sundaramoorthi

Ganesh Sundaramoorthi 

Senior Technical Fellow (Research) in Computer Vision & Machine Learning
Raytheon Technologies (RTX)
Research Center
Email: ganesh.sundaramoorthi ( @ )

About Me

My research lies in both theory and application of artificial intelligence (computer vision and machine learning). My work in these areas develops fundamental mathematical methods at the intersection of optimization, geometry, partial differential equations, and statistics.

Recent News

  • Won an ARPA-E GOPHURRS contract as Co-I to develop 3D neural representations to reconstruct 3D from radar signals, specifically signals from quantum sensors. Excited to collaborate with quantum sensors researchers to co-develop artificial intelligence!

  • New paper on a PDE-based explanation of the Edge of Stability in training deep networks accepted to Journal of Machine Learning Research!

  • Organized a workshop at Raytheon Technolgies on Event-Based Sensing, Algorithms, and Processing Hardware across industry (e.g., Raytheon, Intel, Brainchip), governement (e.g., DARPA, AirForce, SpaceForce) and academia (e.g., Georgia Tech, Purdue, UPenn, UIUC, U. Maryland). Excited for the future of neuromorphic computing in advancing AI!

  • New paper on instabilities in training 3D neural models, and how to address them from a PDE perspective. Accepted to NeurIPS 2023! Code is available.

  • Won an IARPA WRIVA contract. Excited to be leading the Raytheon Technologies team as PI across universities (UT Austin, Georgia Tech, UPenn, Stevens) and industry (Unity, BBN, Adaptive Immersion)! This research seeks to advance 3D vision / generative AI using ground level, aerial, and satellite imagery.

  • New ArXiv Technical Report on translation invariance in CNNs

  • New work (accepted to ICCV 2019) on minimum delay CNN-based object detection. Code available.

  • New ArXiv technical report on discrete analysis of Accelerated PDEs. Accepted to JMIV 2019.

  • Paper on Accelerated PDEs on Manifolds accepted to NIPS 2018

  • My student, Naeemullah Khan, successfully defended his PhD thesis! Nice work on the interface of learning and geometry.

  • Paper on Layered Motion Models accepted to ECCV 2018

  • Invited to present at Ohio State University, AMS Sectional Meeting on Geometric Methods in Shape Analysis (March 2018)

  • Invited to present at Isaac Newton Institute, Cambridge University for workshop on Shape Analysis and Computational Anatomy (Nov 2017)

  • Invited to present seminar at Emory University (Math/CS) department (Nov 2017)

  • Invited to present at Florida State University, Workshop on Applications-Driven Geometric Functional Data Analysis (Oct 2017)

  • Paper on Learning Shape-Tailored Descriptors accepted at CVPR 2018

  • Paper on Minimal Surfaces with Topology accepted at TPAMI (to appear 2018)

  • My student, Marei Algarni, successfully defended his PhD thesis! Nice insights from computational topology for 3D surface segmentation.

  • Invited to serve as Area Chair for CVPR 2018