profile photo

Mayank Mishra

PhD Student
Robotics Institute, Carnegie Mellon University


About Me

I am a second-year PhD student at the Robotics Institute, Carnegie Mellon University, advised by Prof. Sebastian Scherer. My research focuses on robot learning for humanoid robots.

Before starting my PhD, I earned my MS in Robotics [thesis] from CMU, where I was advised by Prof. Srinivasa Narasimhan. During my master's, I was part of CMU’s Team Chiron in the DARPA Triage Challenge. My work focused on developing algorithms for robotic systems to assess human health, in close collaboration with Prof. Laszlo Jeni and Prof. Sebastian Scherer.

Prior to CMU, I was a research assistant at the Indian Institute of Science, Bangalore, working with Prof. Venkatesh Babu. I worked on the generalizability of neural networks under class imbalance and domain shifts. I finished my undergraduate from UPES, Dehradun working with Dr. Tanupriya Choudhury.

(Selected) Publication(s)

  1. Feeling Through the Load arXiv
    Feeling Through the Load: Compliant Quadruped Locomotion under Payload Interactions
    Shaunak A Mehta, Mayank Mishra, Prajit KrisshnaKumar, Sebastian Scherer, & Koichiro Niinuma
    arXiv preprint, 2026

    Enables a quadruped to carry a loose payload and compliantly yield to human pushes on it, using only onboard sensing.

    Paper

  2. CoRL
    Recover, Discover, Plan: Learning Skills and Concepts from Robot Failures
    Bowen Li, Mayank Mishra, Y Isabel Liu, Stone Tao, Nishanth Kumar, Alexander G Gray, Ruwan Wickramarachchi, Jonathan Francis, Sebastian Scherer, & Tom Silver
    Conference on Robot Learning (CoRL 2026)

    Introduces ReSYNC, which learns skills and planning concepts from robot failures to solve unseen long-horizon tasks.

    Paper

  3. DeiT-LT CVPR
    DeiT-LT: Distillation Strikes Back for Vision Transformer Training on Long-Tailed Datasets
    Harsh Rangwani*, Pradipto Mondal*, Mayank Mishra*, Ashish Ramayee Asokan, & R. Venkatesh Babu
    * denotes equal contribution
    Conference on Computer Vision and Pattern Recognition (CVPR 2024)

    Introduces a distillation strategy to train a Vision Transformer from scratch on class imbalanced datasets.

    Paper Code

  4. Saddle-LongTail NeurIPS
    Escaping Saddle Points for Effective Generalisation on Class-Imbalanced Data
    Harsh Rangwani, Sumukh K Aithal, Mayank Mishra, & R. Venkatesh Babu
    Neural Information Processing Systems (NeurIPS 2022)

    Examines the spectral density of Hessian of class wise loss to show network weights converges to a saddle point in the loss landscape of minority classes. Furthermore, shows that methods that encourage convergence to flatter minima can be used to escape saddle points for minority classes.

    Paper Code

  5. SDAT ICML
    A Closer Look at Smoothness in Domain Adversarial Training
    Harsh Rangwani, Sumukh K Aithal, Mayank Mishra, Arihant Jain, & R. Venkatesh Babu
    International Conference on Machine Learning (ICML 2022)

    Analyses the effect of smoothness of loss landscape in domain adversarial training and achieves SOTA results for domain adaptation on Office-Home and VisDA-2017 datasets.

    Paper Code

  6. Allergen30 Journal
    Allergen30: Detecting food items with possible allergens using deep learning based computer vision
    Mayank Mishra, Tanmay Sarkar, Tanupriya Choudhury, Nikunj Bansal, Slim Smaoui, Maksim Rebezov, Mohammad Ali Shariati, & Jose Manuel Lorenzo
    Journal: Food Analytical Methods (SCI, Springer)

    Introduces Allergen30, a custom-made dataset with 6,000+ images of 30 commonly used food items that can trigger an allergic reaction within the human body. This work is one of the first research attempts to train a deep learning based object detection model to detect the presence of such food items from images.

    Paper Dataset


Theme adapted from Minimal Light.