Saket Tiwari
Papers
1
Total Citations
2
H-Index
1
About
Saket Tiwari is a researcher advancing the frontier of meta-reinforcement learning and skill acquisition, with a focus on enabling agents to learn efficiently in complex, long-horizon tasks. His most-cited work, "Meta-Learning Parameterized Skills" (2022), introduces a novel algorithm that learns transferable, parameterized skills and synthesizes them into a compact action space, significantly accelerating learning in challenging environments. By combining off-policy meta-RL with trajectory-centric smoothness constraints, Tiwari’s approach addresses key bottlenecks in skill transfer and generalization. Though early in his career, his contributions have already garnered attention, with his flagship paper accumulating citations that underscore its relevance to the growing field of lifelong learning and robotics. Tiwari’s research bridges the gap between meta-learning and hierarchical control, offering practical pathways for agents to adapt rapidly to new tasks. His work is particularly notable for its potential to reduce sample complexity in real-world applications, from autonomous navigation to robotic manipulation. As a rising voice in AI, Tiwari continues to explore how structured skill representations can unlock more robust and scalable learning systems.
Research Focus
Key Achievements
Top Papers
- 1Meta-Learning Parameterized Skills2 citations · 2022