Arun Venkatraman
Papers
12
Total Citations
557
H-Index
10
About
Arun Venkatraman is a leading researcher at the intersection of robotics, machine learning, and brain-computer interfaces (BCIs). His work focuses on enabling autonomous systems to learn from limited human input, particularly in high-stakes manipulation tasks. Venkatraman’s most influential contribution is the concept of **autonomy-infused teleoperation**, where a robot shares control with a human operator to overcome noisy or low-dimensional commands. This approach proved transformative for BCI-controlled prosthetics, as demonstrated in his highly cited 2016 paper (113 citations) blending brain-machine interfaces with vision-guided robotics to improve grasping performance. His work on **imitation learning** is equally impactful, notably the 2017 paper "Deeply AggreVaTeD" (92 citations), which introduced a differentiable framework for sequential prediction that remains a benchmark in the field. Venkatraman has also advanced autonomous manipulation in unstructured environments, such as pile manipulation using object affordances (87 citations). With over 500 total citations, his research bridges theoretical advances in learning from demonstration with practical, real-world robotic systems, making him a key figure in the push toward more capable and assistive autonomous agents.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5Autonomy Infused Teleoperation with Application to BCI Manipulation46 citations · 2015
- 6
- 7Improved Learning of Dynamics Models for Control31 citations · 2017
- 8Autonomy Infused Teleoperation with Application to BCI Manipulation18 citations · 2015
- 9Feedback in Imitation Learning: The Three Regimes of Covariate Shift17 citations · 2021
- 10Learning to filter with predictive state inference machines11 citations · 2016