Pratik Chaudhari
Associate Professor, ESE
Department: Electrical and Systems Engineering (ESE), Computer and Information Sciences (CIS)
Pratik is interested in the areas of machine learning, robotics and computer vision. His ambition is to bring the dream of cybernetics closer to reality, towards Embodied Intelligence. The ability to perceive and control the environment is the hallmark of such intelligence and their interplay is the core of my research. In a nutshell, he would like to build embodied autonomous agents that are intelligent not just in the sense of a chess-playing program but also intelligent in the sense of a dog. This research is highly multi-disciplinary: it draws from optimization, statistical physics, partial differential equations and stochastic processes, as also, stochastic control, motion planning and formal verification methods. It involves both experimental and theoretical work; pushing the envelope of science and achieving broad societal impact requires both. It focuses on applications in home robotics, personal healthcare and autonomous driving.
Research Areas
- Computer Vision
- Perception
- Reinforcement Learning
- Autonomy
- Machine Learning
- Planning
Publications
Symskill: Symbol and Skill Co-Invention for Data-Efficient and Reactive Long-Horizon Manipulation
Publisher ICRA
Neurosim: A Fast Simulator for Neuromorphic Robot Perception
Publisher CVPR
The LU-Mirage – An independent evaluation of the zero-shot claims in the LUMIR challenge
Publisher MIDL
Analytical characterization of sloppiness in neural networks: Insights from linear models
Publisher Physical Review E
Symskill: Symbol and Skill Co-Invention for Data-Efficient and Real-Time Long-Horizon Manipulation
Publisher CoRL
Fast Feature Field (F3): A Predictive Representation of Events
Publisher arXiv
Toward a science of prospective learning
Publisher Neuron
HALO: High-Altitude Language-Conditioned Monocular Aerial Exploration and Navigation
Publisher ArXiv
SlideSLAM: Sparse, Lightweight, Decentralized Metric-Semantic SLAM for Multirobot Navigation
Publisher IEEE Transactions on Robotics
Asset-Centric Metric-Semantic Maps of Indoor Environments
Publisher ArXiv
Publisher ArXiv
Learning a Vision-Based Footstep Planner for Hierarchical Walking Control
Publisher IEEE-RAS International Conference on Humanoid Robots
Language Modeling with Learned Meta-Tokens
Publisher ICML Workshop LCFM
Prospective Learning in Retrospect
Publisher International Conference on Artificial General Intelligence
Spatio-Temporal Metric-Semantic Mapping for Persistent Orchard Monitoring: Method and Dataset
Publisher IEEE Robotics and Automation Letters
Observability of Latent States in Generative AI Models
Publisher ICLR
An Active Perception Game for Robust Exploration
Publisher ICRA
Gaussian Splatting as a Unified Representation for Autonomy in Unstructured Environments
Publisher ArXiv
From Linearity to Non-Linearity: How Masked Autoencoders Capture Spatial Correlations
Publisher ICCV
Design and Evaluation of Motion Planners for Quadrotors in Environments with Varying Complexities
Publisher ICRA
Publisher HRI Workshop
Publisher Generative Modeling meets HRI - RSS'24 Workshop
Prospective Learning: Learning for a Dynamic Future
Publisher NeurIPS
Heat death of generative models in closed-loop learning
Publisher CDC
Deep Learning in Medical Image Registration: Magic or Mirage?
Publisher NeurIPS
Generative models of MRI-derived neuroimaging features and associated dataset of 18,000 samples
Publisher Scientific Data
Active Scout: Multi-Target Tracking Using Neural Radiance Fields in Dense Urban Environments
Publisher IROS
Publisher ArXiv
A Confidence Interval for the ℓ2 Expected Calibration Error
Publisher ArXiv
Active Perception using Neural Radiance Fields
Publisher ACC
Meanings and Feelings of Large Language Models: Observability of Latent States in Generative AI
Publisher ArXiv
From NeRFs to Gaussian Splats, and Back
Publisher ArXiv
Publisher Proceedings of SPIE--the International Society for Optical Engineering
Automated estimation of microcirculation capillary density using relative perfusion maps
Publisher Proc. SPIE 12926, Medical Imaging 2024: Image Processing
TreeScope: An Agricultural Robotics Dataset for LiDAR-Based Mapping of Trees in Forests and Orchards
Publisher ICRA
Publisher IEEE International Conference on Robotics and Automation (ICRA)
Scalable Reinforcement Learning Policies for Multi-Agent Control
Publisher arXiv
PennSyn2Real: Training Object Recognition Models Without Human Labeling
Publisher IEEE Robotics and Automation Letters
A Free-Energy Principle for Representation Learning
Publisher IEEE International Conference on Robotics and Automation (ICRA)
A Geometric Interpretation of Stochastic Gradient Descent Using Diffusion Metrics
Publisher Multidisciplinary Digital Publishing Institute (MDPI)
BayesRace: Learning to Race Autonomously from prior experience
Publisher arXiv
DIRECTIONAL ADVERSARIAL TRAINING FOR COST SENSITIVE DEEP LEARNING CLASSIFICATION APPLICATIONS
Publisher arXiv
Rethinking the Hyperparameters of Fine-tuning
Publisher IEEE International Conference on Robotics and Automation (ICRA)
TraDE: Transformers for Density Estimation
Publisher arXiv
Proximal Deterministic Policy Gradient
Publisher IEEE International Conference on Intelligent Robots and Systems (IROS)
A Baseline for Few-Shot Image Classification
Publisher IEEE International Conference on Robotics and Automation (ICRA)
Publisher IEEE International Conference on Robotics and Automation (ICRA)
P30: Policy-on Policy-off Policy Optimization
Publisher arXiv
News
Student News
When Robots Dive into Forests: Exploring Autonomous Systems in Natural Environments
September 21, 2024
Faculty News
Building Robots that Learn: Pratik Chaudhari Investigates the Underpinnings of AI
April 5, 2024
Faculty News
Intel® Rising Star Faculty Award 2022 Recognizes Research Efforts of Dr. Pratik Chaudhari
August 18, 2023
Faculty News
Pratik Chaudhari receives a 2019 Q4 of AWS Machine Learning Research Award
July 6, 2020
Ph.D. Students
Research Staff
Alumni
Postdoctoral Researcher, BE '22-'25 - Assistant Professor, New Jersey Institute of Technology
Research Areas
- Computer Vision
- Perception
- Reinforcement Learning
- Autonomy
- Machine Learning
- Planning