Hierarchical Interest-Driven Goal Babbling for Efficient Bootstrapping of Sensorimotor skills

Sensorimotor Self-organization via Circular-Reactions

“…In particular, after going through the entire learning session, under the condition without noise, the median Cartesian error is 0.16 cm, and the third quarter error value is 4.96 cm. These results indicate that the position controller is able to deliver the end-effector to contact or reach the target with satisfactory error levels, compared to other recent studies (Mahoor et al, 2016 ; Nguyen et al, 2019 ; Rayyes et al, 2020 ). As an example, Mahoor reported a median Euclidean distance error of approximately 4 cm achieved by neural-networks learned through motor babbling.…”

Intrinsic motivation learning for real robot applications

“…The competence-based signal is the forgetting factor which monitors where the robot's performance deteriorates during lifelong learning. This combination of different learning signals led to high sampleefficiency which facilitates online data-driven direct learning on real robots without any pre-training in simulation as shown in (Rayyes et al, 2020a;Rayyes et al, 2020b;Rayyes, 2020).…”

“…However, an experiment in (Baranes et al, 2014) showed that humans learn by maximizing their knowledge of a task and their competence. Accordingly, a recent intrinsic motivation method named "Interest Measurement" (Rayyes et al, 2020b) combined both knowledge-based and competence-based signals.…”

“…While most intrinsic motivation methods in the literature are either knowledge-based or competence-based, an experimental study (Baranes et al, 2014) showed that humans tend to learn by trying to improve their knowledge about the tasks in hand and their competence. Interest Measurement (Rayyes et al, 2020b) is a recent intrinsic motivation method which combines knowledge-based with competence-based signals. The knowledge-based signal, named relative error, drives the robot's exploration toward difficult to attain goals/tasks, e.g., goals near the border of the robot's workspace.…”

“…In my opinion, increasing the sample-efficiency and the applicability of the intrinsic motivation methods can be done by combining them with mental replay (Andrychowicz et al, 2017;Rayyes et al, 2020b) and goal-directed methods, e.g., Goal Babbling (Rolf et al, 2011), as shown in the literature so far.…”