Minttu Alakuijala

About

Minttu Alakuijala is an emerging researcher specializing in robotic learning, reinforcement learning, and human-robot knowledge transfer. Her work sits at the intersection of imitation learning and reinforcement learning, with a particular focus on enabling robots to acquire skills more efficiently and with minimal human engineering effort. Her most notable contribution, "Learning Reward Functions for Robotic Manipulation by Observing Humans" (2023, 16 citations), addresses one of robotics' core challenges: bridging the gap between human demonstration and robotic execution. By leveraging video observations of human manipulators, her approach offers a scalable, cost-effective pathway to training robotic policies without requiring costly teleoperation or manual reward engineering. This work reflects a growing recognition that rich, naturalistic human data can serve as a powerful supervisory signal. Complementing this, her earlier work on "Residual Reinforcement Learning from Demonstrations" (2021, 11 citations) extends the residual RL framework to handle visual inputs and sparse rewards — notoriously difficult conditions — by grounding learning in demonstrations. Together, these contributions position Alakuijala as a thoughtful contributor to the challenge of making robotic manipulation more accessible, data-efficient, and grounded in natural human behavior.

Research Focus

Key Achievements

Top Papers

1. 1
   Learning Reward Functions for Robotic Manipulation by Observing Humans

   16 citations · 2023
2. 2
   Residual Reinforcement Learning from Demonstrations

   11 citations · 2021

Key Collaborators

Contact & Links