Papers
2
Total Citations
24
H-Index
2
About
Kirill Aksenov is a researcher advancing the frontiers of deep reinforcement learning (RL), with a focus on making it more sample-efficient and computationally practical. His key contributions center on hierarchical reinforcement learning, where he develops methods to break complex tasks into manageable subtasks, significantly reducing the number of interactions an agent needs with its environment. Aksenov’s most influential work, “Forgetful Experience Replay in Hierarchical Reinforcement Learning from Expert Demonstrations” (2021), has garnered 22 citations, reflecting its impact on the field. In this paper, he introduces a novel “forgetful” mechanism for experience replay that intelligently prioritizes and discards past experiences, allowing agents to learn more effectively from limited expert demonstrations. This approach tackles a critical bottleneck in RL: the enormous computational cost and data hunger of traditional methods. By combining hierarchical learning with selective memory, Aksenov’s work paves the way for more efficient training in complex domains like gaming and robotics. His research is particularly valuable for students and practitioners seeking to deploy RL in real-world settings where data and compute resources are constrained.
Research Focus
Key Achievements
Top Papers
- 1
- 2