Ryuki Tachibana

IBM Research - Tokyo

Papers

6

Total Citations

474

H-Index

6

About

Ryuki Tachibana is a leading researcher at the intersection of deep reinforcement learning and high-precision industrial robotics. His work directly tackles one of manufacturing’s most stubborn challenges: enabling robots to perform assembly tasks with accuracy that exceeds their own mechanical precision. Tachibana’s most influential contribution, “Deep reinforcement learning for high precision assembly tasks” (300 citations), demonstrates how a robot can master the classic peg-in-hole problem through trial-and-error learning, eliminating the need for tedious manual parameter tuning. To make such learning viable in the real world, he developed OptLayer, a practical constrained optimization framework for deep RL (118 citations) that prevents the unsafe, unconstrained motions typical of simulated agents. This innovation bridges the gap between simulated training and real-world deployment. His work also extends to human-like motion prediction using Long Short-Term Memory networks and the creation of experimental force-torque datasets for multi-shape insertion tasks. By systematically replacing hand-crafted control logic with learned policies, Tachibana is paving the way for flexible, autonomous manufacturing systems that can adapt to new tasks without extensive reprogramming.

Research Focus

Key Achievements

6
H-Index
6
Papers
474
Total Citations
79
Avg Citations/Paper
🏆 Most Cited Paper
Deep reinforcement learning for high precision assembly tasks
300 citations · 2017
📈 Most Prolific Year: 2017 (4 Papers)
🤝 Key Collaborators: 9
🏛 Institutions: IBM Research - Tokyo

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 14 days ago