On Learning, Representing, and Generalizing a Task in a Humanoid Robot

By Sylvain Calinon, Florent Guenter, Aude Gemma Billard. 2007

1,243 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Control and Systems Engineering, Robotics & Control

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Control and Systems Engineering

In Robotics & Control

Citations in the corpus

Benchmark for Skill Learning from Demonstration: Impact of User Experience, Task Complexity, and Start Configuration on Performance (2019, background): “Based on the choice of motion generation strategy, the majority of skill learning approaches can be further broadly categorized into four groups: time-invariant (stable) dynamical systems [2], [3], [13], time-dependent dynamical systems [7], [9], [10], time-dependent statistical methods [8], [14], [”
Third-Person Imitation Learning (2017, background): “Such a mapping is often difﬁcult to obtain, and it typically relies on providing feature representations that captures the invariance between both environments Carpenter et al. (2002); Shon et al. (2005); Calinon et al. (2007); Nehaniv (2007); Gioioso et al. (2013); Gupta et al. (2016).”
An Algorithmic Perspective on Imitation Learning (2018, methodology): “A policy for planning a desired trajectory can be often learned as an open-loop policy since it can be addressed as a one shot decision making for a given situation such as in [Calinon et al., 2007, Takano and Nakamura, 2015].”
Vision-Based Multi-Task Manipulation for Inexpensive Robots Using End-to-End Learning from Demonstration (2017, methodology): “One possible technique is to hand-engineer task-specific features [13], [14].”
BM: An iterative algorithm to learn stable non-linear dynamical systems with Gaussian mixture models (2010, background): “Programming by Demonstrations (PbD) can be used to learn such motion primitives from a few demonstrations1 of the motions performed by a trained agent (human or robot) [1], [3]–[5].”
Probabilistic model-based imitation learning (2013, background): “, at the symbolic [43] or trajectory level [10]), whether they use a system model, and whether/how they employ cost functions for the task.”
A case study of trajectory transfer through non-rigid registration for a simplified suturing scenario (2013, background): “A closely related approach is to fit a Gaussian (or mixture of Gaussians) distribution to the collection of demonstrations and identify the directions of low variance– these are the important directions [3], [14], [2].”
Affordances from Human Videos as a Versatile Representation for Robotics (2023, background): “Extensive work has been done on Learning from Demonstrations (LfD) where human supervision is usually provided through teleoperation (of a joystick or VR interface) [77, 115, 133] or kinesthetic teaching, where a user physically moves the robot arm [13, 16, 27, 70, 94].”
Learning compact parameterized skills with a single regression (2013, methodology): “More sophisticated methods may be used to model this variation [2], but this is beyond the scope of this paper.”
Autonomous Helicopter Aerobatics through Apprenticeship Learning (2010, methodology): “…apprenticeship learning algorithms for modeling the dynamics of the helicopter, (iii) combining these apprenticeship learning algorithms with (a variation of) existing optimal control algorithms, namely, a receding horizon variation of linear quadratic control methods for non-linear systems (a…”
Novel Interaction Strategies for Learning from Teleoperation (2012, methodology): “It has been previously used in similar LfD scenarios (Calinon, Guenter, and Billard 2007) and found to be very useful.”
BatchBALD: Efficient and Diverse Batch Acquisition for Deep Bayesian Active Learning (2019, background): “AL has already made real-world impact in manufacturing [34], robotics [5], recommender systems [1], medical imaging [18], and NLP [31], motivating the need for pushing AL even further.”
R-IAC: Robust Intrinsically Motivated Exploration and Active Learning (2009, methodology): “Another approach, based on Gaussian mixture regression [3], [49], is based on the learning of the joint probability distribution of the sensorimotor variables, instead of learning a forward or an inverse model, and can be used online for inferring specific forward or inverse models by well chosen pr”
Keyframe-based Learning from Demonstration (2012, methodology): “Dynamical system approaches such as Stable Es-timator of Dynamical Systems (SEDS) [17] and Dynamic Movement Primitives (DMP) [24] as well as mixture models ( e.g. Gaussian Mixture Models as in [10]) are policy learning methods.”