Skand

skand.jpg

I am a postdoc at UT Austin 🤘 working with Roberto Martín-Martín in the RobIN lab. Previously, I was visiting the WEIRD lab at the University of Washington 🐺, collaborating with Abhishek Gupta and Jesse Zhang on fast real-world adaptation of policies using world models.

I received my Ph.D. from Oregon State University 🦫, where I was advised by Fuxin Li and Stefan Lee. My dissertation, Efficient and Robust Robot Learning, proposed methods to build robots that keep working – making them (a) more energy-efficient, (b) resilient to sensory failures, (c) robust to viewpoint and lighting changes, and (d) adaptable to new tasks via 3D point dynamics models.

I am interested in developing systems and algorithms that enable robots to learn and continuously adapt to complex tasks in the real world.

Previous Research Life

Previously, I worked with Sungjin Ahn on object-centric world models and RL. Before that, I was worked with Venkatesh Babu at the Indian Institute of Science. I graduated with Bachelors of Technology (B.Tech) degree in Computer Science and Engineering from Indian Institute of Technology, Ropar where I was advised by Narayanan Krishnan for my Bachelor’s Thesis. During my undergrad, I also worked with Deepti Bathula and Abhinav Dhall.

selected publications

  1. CoRLOral
    Non-conflicting Energy Minimization in Reinforcement Learning based Robot Control
    Skand PeriAkhil Perincherry*Bikram Pandit*, and 1 more author
    Conference on Robot Learning, 2025
  2. CoRL
    Simple Masked Training Strategies Yield Control Policies That Are Robust to Sensor Failure
    Skand PeriBikram Pandit, Chanho Kim, and 2 more authors
    Conference on Robot Learning, 2024
  3. ICRA
    Point Cloud Models Improve Visual Robustness in Robotic Learners
    Skand Peri, Iain Lee, Chanho Kim, and 3 more authors
    International Conference on Robotics and Automation, 2024
  4. ICLR
    SPACE: Unsupervised Object-Oriented Scene Representation via Spatial Attention and Decomposition
    Zhixuan Lin*, Yi-Fu Wu*, Skand Peri*, and 5 more authors
    International Conference on Learning Representations, 2020