Papers
36
Total Citations
1,099
H-Index
19
About
Daniel Seita is a robotics researcher whose work centers on robotic manipulation of deformable objects, imitation and reinforcement learning, and surgical robotics automation. His most influential contributions tackle one of robotics' most persistent challenges: enabling robots to reliably handle flexible, high-dimensional materials such as cables, fabrics, and bags. His 2021 paper on goal-conditioned Transporter Networks for deformable object rearrangement (122 citations) and his work on deep imitation learning for fabric smoothing (109 citations) have become foundational references in the field, demonstrating that learning-based approaches can overcome the complex dynamics that stymied classical methods. His research on sim-to-real transfer, visual foresight, and dense visual correspondences further advances robots' ability to generalize manipulation skills across diverse tasks without exhaustive real-world training data. Beyond soft-body manipulation, Seita has made notable contributions to surgical robotics, developing fast and reliable debridement systems for cable-driven surgical robots and leveraging depth-sensing to automate delicate clinical subtasks. His foray into risk-averse adversarial reinforcement learning (64 citations) reflects a broader commitment to robust, safe policy learning. With over 600 cumulative citations, Seita's research bridges fundamental learning theory and practical robotic dexterity, making him a prominent voice in modern manipulation research.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Deep Transfer Learning of Pick Points on Fabric for Robot Bed-Making81 citations · 2022
- 4
- 5VisuoSpatial Foresight for Multi-Step, Multi-Task Fabric Manipulation71 citations · 2020
- 6Risk Averse Robust Adversarial Reinforcement Learning64 citations · 2019
- 7
- 8VisuoSpatial Foresight for physical sequential fabric manipulation41 citations · 2021
- 9
- 10