Varan R. Kompella

Dalle Molle Institute for Artificial Intelligence Research

Papers

Total Citations

H-Index

About

Varan R. Kompella is a leading researcher in developmental robotics and autonomous skill acquisition, with a focus on enabling robots to learn complex sensorimotor behaviors from raw sensory input. His most cited work, "Explore to see, learn to perceive, get the actions for free: SKILLABILITY" (2014, 11 citations), introduces a groundbreaking framework that allows humanoid robots to autonomously discover and refine multiple skills through curiosity-driven exploration, without requiring pre-programmed knowledge. By integrating hierarchical reinforcement learning with intrinsic motivation, Kompella’s approach enables robots to learn perceptual and motor abilities directly from high-dimensional visual streams—a significant departure from traditional supervised methods. This work has influenced the field of open-ended learning, demonstrating how robots can bootstrap their own development by treating exploration as a primary driver. Kompella’s contributions are particularly notable for bridging the gap between unsupervised learning and real-world robotic control, offering a scalable path toward more adaptive and autonomous systems. His research continues to inspire new directions in lifelong learning for artificial agents, where machines not only act but actively seek out new challenges to master.