Associate Professor · Georgia Tech

I lead the Robotics Perception and Learning (RIPL) lab at Georgia Tech's School of Interactive Computing and serve as Associate Director of ML@GT. We work on generalization and robustness at the intersection of machine learning/artificial intelligence, perception, and robotics.

  • CODA S1181B
  • zkira at gatech dot edu

Talks & podcasts

Older talks (2023 & earlier)
  • Video Slides CoRL 2023 LangRob — Act, Interact, and Finetune
  • Video CVPR 2023 CLVISION — Continual Fine-tuning of Foundation Models

Latest news

What's new in the lab

Year highlights (2024 & earlier)

2024

We had 2 ICRA, 1 ICLR, 3 CVPR, 2 ECCV, 3 NeurIPS, and 1 TMLR papers, spanning vision-language-action models, open-world perception and navigation, neural fields, and robust finetuning of Foundation Models. Promoted to Associate Professor in August. Invited talks at the ECCV OOD-CV and FOCUS workshops. Co-organized the CVPR workshop on 3D Vision-Language Models and RoboNerF, the 1st Workshop on Neural Fields in Robotics. James Smith won the Outstanding GRA Award and Ram Ramrakhya the CoC Rising Star Doctoral Student Research Award. Graduated Junjiao Tian (Ph.D.).

2023

We had 1 ICRA, 1 ICLR, 4 CVPR, 1 ICML, 1 ICCV, 1 CoRL, 3 NeurIPS, and 2 WACV papers. Invited talks at the CVPR 2023 CLVISION workshop, CVIT Summer School on AI, and CoRL 2023 LangRob workshop. Co-organized the ICCV tutorial on Continual Learning and the NeurIPS HomeRobot Challenge. Area Chair for CVPR, ICLR, NeurIPS, and ICRA. Received the NSF CAREER Award, the College of Computing Outstanding Junior Faculty Research Award, and the EURASIP Best Paper Award. James Smith was accepted to the CVPR Doctoral Colloquium and Nathan Glaser won 2nd place paper at the ICRA CoPerception Workshop. Graduated 5 Ph.D. students: James Smith (Samsung), Nathan Glaser (Zoox), Yen-Cheng Liu (Meta), Zubair Irshad (TRI), and Chia-Wen Kuo (TikTok).

2022

We had 2 ICRA, 2 CVPR, 1 Nature Machine Intelligence, 2 ECCV, and 1 NeurIPS papers. Invited talks at UIUC, Vanderbilt, and an ECCV workshop. Co-organized the 2nd workshop on Learning from Limited and Imperfect Data (L2ID). Funding from IRIM/IPaT, TRI, and Google. Andrew Szot won the Outstanding Online Teaching Assistant of the Year Award.

2021

We had 1 ICLR, 2 ICRA, 1 ICCV, 2 NeurIPS spotlight papers, and 1 IJCNN paper. Area Chair for ICLR and NeurIPS. Significant press for Habitat 2.0. Funding for DARPA LwLL and DARPA L2M Phase II projects. Invited talks at Google and Microsoft AI, and co-organized the CVPR L2ID Workshop.

2020

We had 1 AAAI, 2 ICRA, 3 CVPR, and 2 ECCV papers. Partnered with Facebook on a co-teaching program, served as Area Chair for NeurIPS, and gave invited talks at the VL3, Agriculture Vision, and ULAD-2020 workshops.

2019

We had 3 ICLR, 1 ICRA, 1 CVPR, 1 Oral ICCV, 2 journal, 1 WACV, and ICLR/IROS workshop papers. Received funding from DARPA LwLL and Samsung.

2018

We had ICLR, CVPR, WACV, NeurIPS Continual Learning Workshop, and journal papers. New funding from DARPA L2M and ONR. 08/2018 — Assistant Professor!

About

Bio

I am an Associate Professor at the School of Interactive Computing in the College of Computing, and serve as an Associate Director of ML@GT, the machine learning center at Georgia Tech. Previously I was a Branch Chief at the Georgia Tech Research Institute (GTRI) and a Research Scientist at SRI International Sarnoff in Princeton. I received my Ph.D. in 2010 with Professor Ron Arkin as my advisor.

I lead the Robotics Perception and Learning (RIPL) lab. Our work lies at the intersection of machine learning and artificial intelligence for perception and robotics, focusing on generalization and robustness. Recent works include robust finetuning of vision-language models to preserve out-of-distribution generalization capabilities, open-world generalization, long-horizon RL, 3D processing, and fine-tuning of multi-modal Foundation Models into Vision-Language-Action models via supervised finetuning, reinforcement learning, and post-training. I have grown a portfolio of projects funded by NSF, ONR, DARPA, and industry (Samsung, TRI, Google, and Meta), and have won the NSF CAREER Award, the College of Computing Outstanding Junior Faculty Research Award, and several best paper/student paper awards.

RIPL — Robotics Perception and Learning lab logo

Machine learning for perception and robotics: generalization, robustness, and embodied AI.

Meet the group

Interested in joining RIPL?

I am always looking for strong Ph.D. students and collaborators in machine learning, perception, and robotics. See the join page for current openings and how to apply.