Learning From Demonstration

Learning from Demonstration (LfD) focuses on enabling robots and other systems to acquire skills by observing and imitating human demonstrations, aiming to reduce the need for complex manual programming. Current research emphasizes improving LfD's efficiency and generalization capabilities, exploring techniques like curriculum learning, hypernetworks for continual learning, and the integration of formal specifications (e.g., temporal logic) to ensure robust performance. This field is significant for advancing robotics, automation, and AI, offering more intuitive and data-efficient methods for training intelligent agents to perform complex tasks in various domains.

Papers

June 11, 2024

A Practical Roadmap to Learning from Demonstration for Robotic Manipulators in Manufacturing
Alireza Barekatain, Hamed Habibi, Holger Voos
LeArning Abstract Robotic Manipulator Textual Demonstration Manufacturing Industry Requirement Based Roadmap Learning From Demonstration Industrial Task Industrial Manipulator

May 6, 2024

Reverse Forward Curriculum Learning for Extreme Sample and Demonstration Efficiency in Reinforcement Learning
Stone Tao, Arth Shukla, Tse-kai Chan, Hao Su
Reinforcement Learning Evidence Piece Curriculum Reinforcement Learning Demonstration Data Learning From Demonstration Sample Quality

March 30, 2024

Synthetic Dataset Generation and Learning From Demonstration Applied to Industrial Manipulation
Alireza Barekatain, Hamed Rahimi Nohooji, Holger Voos
LeArning Abstract Synthetic Data Generation Textual Demonstration Robot Skill Robot Programming Learning From Demonstration Industrial Task Manipulation Pipeline

November 6, 2023

Scalable and Efficient Continual Learning from Demonstration via a Hypernetwork-generated Stable Dynamics Model
Sayantan Auddy, Jakob Hollenstein, Matteo Saveriano, Antonio Rodríguez-Sánchez, Justus Piater
Potential Scalability Learning Dynamic Textual Demonstration Efficient Continual Learning Learning From Demonstration

October 15, 2023

Auto-LfD: Towards Closing the Loop for Learning from Demonstrations
Shaokang Wu, Yijin Wang, Yanlong Huang
LeArning Abstract Generalization Performance Noisy Demonstration Dynamic Movement Primitive Learning From Demonstration Peg in Hole Task Control Input Inference

May 4, 2023

Programming-by-Demonstration for Long-Horizon Robot Tasks
Noah Patton, Kia Rahmani, Meghana Missula, Joydeep Biswas, Işil Dillig
Long Horizon Learning From Demonstration Control Flow Programmatic Reinforcement Learning Programming by Example

March 24, 2023

Learning from Few Demonstrations with Frame-Weighted Motion Generation
Jianyong Sun, Jens Kober, Michael Gienger, Jihong Zhu
Motion Generation Robot Motion Robot Programming Robot Environment Learning From Demonstration

November 28, 2022

Autonomous Assessment of Demonstration Sufficiency via Bayesian Inverse Reinforcement Learning
Tu Trinh, Haoyu Chen, Daniel S. Brown
Inverse Reinforcement Learning Evidence Piece Learning From Demonstration Optimal Demonstration

September 27, 2022

Follow The Rules: Online Signal Temporal Logic Tree Search for Guided Imitation Learning in Stochastic Domains
Jasmine Jerry Aloor, Jay Patrikar, Parv Kapoor, Jean Oh, Sebastian Scherer
Monte Carlo Tree Search Signal Temporal Logic Simple RULE Learning From Demonstration

September 7, 2022

Optimizing Demonstrated Robot Manipulation Skills for Temporal Logic Constraints
Akshay Dhonthi, Philipp Schillinger, Leonel Rozo, Daniele Nardi
Manipulation Task Signal Temporal Logic Expert Demonstration Manipulation Skill Formal Specification Learning From Demonstration

August 31, 2022

Let Me Check the Examples: Enhancing Demonstration Learning via Explicit Imitation
Sirui Wang, Kaiwen Wei, Hongzhi Zhang, Yuntao Li, Wei Wu
Fewer Example Contrastive Training Learning From Demonstration

August 3, 2022

Robot Learning from Demonstration Using Elastic Maps
Brendan Hertel, Matthew Pelland, S. Reza Ahmadzadeh
Robot Learning Robotic Task Textual Demonstration Robot Skill Learning From Demonstration Elastic Band Provided Demonstration

June 13, 2022

Silver-Bullet-3D at ManiSkill 2021: Learning-from-Demonstrations and Heuristic Rule-based Methods for Object Manipulation
Yingwei Pan, Yehao Li, Yiheng Zhang, Qi Cai, Fuchen Long, Zhaofan Qiu, Ting Yao, Tao Mei
Imitation Learning Robotic Arm Object Manipulation Learning From Demonstration Demonstration Trajectory Stable Manipulation

May 10, 2022

Exploring Viable Algorithmic Options for Learning from Demonstration (LfD): A Parameterized Complexity Approach
Todd Wareham
LeArning Abstract Polynomial Time Textual Demonstration Parameterized Complexity Learning From Demonstration

March 21, 2022

Self-Imitation Learning from Demonstrations
Georgiy Pshikhachev, Dmitry Ivanov, Vladimir Egorov, Aleksei Shpilman
Reinforcement Learning Sparse Reward Noisy Demonstration Suboptimal Demonstration Learning From Demonstration

Learning From Demonstration

Papers

A Practical Roadmap to Learning from Demonstration for Robotic Manipulators in Manufacturing

Reverse Forward Curriculum Learning for Extreme Sample and Demonstration Efficiency in Reinforcement Learning

Synthetic Dataset Generation and Learning From Demonstration Applied to Industrial Manipulation

Scalable and Efficient Continual Learning from Demonstration via a Hypernetwork-generated Stable Dynamics Model

Auto-LfD: Towards Closing the Loop for Learning from Demonstrations

Programming-by-Demonstration for Long-Horizon Robot Tasks

Learning from Few Demonstrations with Frame-Weighted Motion Generation

Autonomous Assessment of Demonstration Sufficiency via Bayesian Inverse Reinforcement Learning

Follow The Rules: Online Signal Temporal Logic Tree Search for Guided Imitation Learning in Stochastic Domains

Optimizing Demonstrated Robot Manipulation Skills for Temporal Logic Constraints

Let Me Check the Examples: Enhancing Demonstration Learning via Explicit Imitation

Robot Learning from Demonstration Using Elastic Maps

Silver-Bullet-3D at ManiSkill 2021: Learning-from-Demonstrations and Heuristic Rule-based Methods for Object Manipulation

Exploring Viable Algorithmic Options for Learning from Demonstration (LfD): A Parameterized Complexity Approach

Self-Imitation Learning from Demonstrations