Noisy Demonstration

Noisy demonstration research focuses on improving machine learning models' ability to learn effectively from imperfect or incomplete training data, a common challenge in real-world applications. Current research emphasizes developing algorithms and model architectures that can filter noise, infer underlying patterns from diverse or suboptimal demonstrations, and leverage techniques like inverse reinforcement learning, behavior cloning, and various forms of few-shot learning to enhance performance. This work is significant because it addresses a critical limitation in many machine learning approaches, paving the way for more robust and reliable systems in fields ranging from robotics and natural language processing to autonomous driving and healthcare.

Papers

May 19, 2023

Characterizing tradeoffs between teaching via language and demonstrations in multi-agent systems
Dhara Yu, Noah D. Goodman, Jesse Mu
Strong Generalization Human Language Multi Agent System Noisy Demonstration Neural Agent Human Communication Communication Round

May 8, 2023

DEFENDER: DTW-Based Episode Filtering Using Demonstrations for Enhancing RL Safety
André Correia, Luís Alexandre
Noisy Demonstration Task Agnostic RL Agent Multiple Defender Unsupervised Episode

April 27, 2023

Inferring Preferences from Demonstrations in Multi-objective Reinforcement Learning: A Dynamic Weight-based Approach
Junlin Lu, Patrick Mannion, Karl Mason
Noisy Demonstration Multi Objective Reinforcement Learning Multi Objective Markov Decision Process Preference Inference

March 28, 2023

March 23, 2023

Boosting Reinforcement Learning and Planning with Demonstrations: A Survey
Tongzhou Mu, Hao Su
Reinforcement Learning Timely Survey Task Planning Noisy Demonstration Sequential Decision Making Robotics Benchmark Trial and Error

March 14, 2023

How Many Demonstrations Do You Need for In-context Learning?
Jiuhai Chen, Lichang Chen, Chen Zhu, Tianyi Zhou
Large Language Model Context Learning Complex Reasoning Noisy Demonstration Reasoning ChAin

March 9, 2023

March 2, 2023

February 15, 2023

When Demonstrations Meet Generative World Models: A Maximum Likelihood Framework for Offline Inverse Reinforcement Learning
Siliang Zeng, Chenliang Li, Alfredo Garcia, Mingyi Hong
Autonomous Driving Estimation Task Inverse Reinforcement Learning Noisy Demonstration Continuous Control Task Generative World Model

February 7, 2023

Leveraging Demonstrations to Improve Online Learning: Quality Matters
Botao Hao, Rahul Jain, Tor Lattimore, Benjamin Van Roy, Zheng Wen
Multi Armed Bandit Online Learning Quality Issue Noisy Demonstration Thompson Sampling Prior Dependent

December 13, 2022

Diverse Demonstrations Improve In-context Compositional Generalization
Itay Levy, Ben Bogin, Jonathan Berant
Context Learning Compositional Generalization Noisy Demonstration

December 12, 2022

MoDem: Accelerating Visual Model-Based Reinforcement Learning with Demonstrations
Nicklas Hansen, Yixin Lin, Hao Su, Xiaolong Wang, Vikash Kumar, Aravind Rajeswaran
Deep Reinforcement Learning Noisy Demonstration Model Learning Visuo Motor Control Visual Model Based

December 7, 2022

Accelerating Self-Imitation Learning from Demonstrations via Policy Constraints and Q-Ensemble
Chao Li
Reinforcement Learning Deep Reinforcement Learning Noisy Demonstration Online Reinforcement Learning Expert Demonstration Policy Constraint

December 3, 2022

December 2, 2022

Prim-LAfD: A Framework to Learn and Adapt Primitive-Based Skills from Demonstrations for Insertion Tasks
Zheng Wu, Wenzhao Lian, Changhao Wang, Mengxi Li, Stefan Schaal, Masayoshi Tomizuka
Reinforcement Learning New Framework Noisy Demonstration Insertion Task Test Point Insertion

October 31, 2022

Learning Modular Robot Locomotion from Demonstrations
Julian Whitman, Howie Choset
Robot Person Policy Learning Noisy Demonstration Modular System Modular Robot Policy Reproducibility

Noisy Demonstration

Papers

Characterizing tradeoffs between teaching via language and demonstrations in multi-agent systems

DEFENDER: DTW-Based Episode Filtering Using Demonstrations for Enhancing RL Safety

Inferring Preferences from Demonstrations in Multi-objective Reinforcement Learning: A Dynamic Weight-based Approach

BC-IRL: Learning Generalizable Reward Functions from Demonstrations

Coordinated Multi-Robot Shared Autonomy Based on Scheduling and Demonstrations

Boosting Reinforcement Learning and Planning with Demonstrations: A Survey

How Many Demonstrations Do You Need for In-context Learning?

Learning Rational Subgoals from Demonstrations and Instructions

ICL-D3IE: In-Context Learning with Diverse Demonstrations Updating for Document Information Extraction

Teach a Robot to FISH: Versatile Imitation from One Minute of Demonstrations

Programmatic Imitation Learning from Unlabeled and Noisy Demonstrations

When Demonstrations Meet Generative World Models: A Maximum Likelihood Framework for Offline Inverse Reinforcement Learning

Leveraging Demonstrations to Improve Online Learning: Quality Matters

Diverse Demonstrations Improve In-context Compositional Generalization

MoDem: Accelerating Visual Model-Based Reinforcement Learning with Demonstrations

Accelerating Self-Imitation Learning from Demonstrations via Policy Constraints and Q-Ensemble

Can In-context Learners Learn a Reasoning Concept from Demonstrations?

Reinforcement learning with Demonstrations from Mismatched Task under Sparse Reward

Prim-LAfD: A Framework to Learn and Adapt Primitive-Based Skills from Demonstrations for Insertion Tasks

Learning Modular Robot Locomotion from Demonstrations