Towards Closing the Loop: Active Learning for Robotics - RSS'10 Workshop
The ability to adapt to changing environment autonomously will be essential for future robots. While this need is well-recognized, most machine learning research focuses largely on perception and static data sets. Instead, future robots need to interact with the environment to generate the data that is needed to foster real-time adaptation based on all information collected in previous interactions and observations. In other words, we need to close the loop between the robot acting, robot sensing and robot learning. Novel active methods need to outperform passive methods by a margin that compensates the potential the extra computational burden and the cost of the active data sampling.
During the last years, there has been an increasing interest in related techniques that could potentially become applicable in this context. These include techniques from statistics such as adaptive sensing or sequential experimental design as well novel reinforcement learning methods that have the potential to scale into robotics. In this context, we would like to bring together researchers from both the robotics and active machine learning in order to discuss for which problems the autonomous learning loop can be closed using learning, and to identify the machine learning methods that can be used to close it.
Detailed information can be found at RSS 2010 Workshop on Active Learning for Robotics.
Disclaimer: VideoLectures.NET emphasizes we are not the authors of these recordings and that the overall quality of these videos was notably improved.