paper-with-me

홈 › Papers

Self-Supervised State-Control through Intrinsic Mutual Information Rewards

2019-09-25 · Rui Zhao, Volker Tresp, Wei Xu

Learning to discover useful skills without a manually-designed reward function would have many applications, yet is still a challenge for reinforcement learning. In this paper, we propose Mutual Information-based State-Control (MISC), a new self-supervised Reinforcement Learning approach for learning to control states of interest without any external reward function. We formulate the intrinsic objective as rewarding the skills that maximize the mutual information between the context states and the states of interest. For example, in robotic manipulation tasks, the context states are the robot states and the states of interest are the states of an object. We evaluate our approach for different simulated robotic manipulation tasks from OpenAI Gym. We show that our method is able to learn to manipulate the object, such as pushing and picking up, purely based on the intrinsic mutual information rewards. Furthermore, the pre-trained policy and mutual information discriminator can be used to accelerate learning to achieve high task rewards. Our results show that the mutual information between the context states and the states of interest can be an effective ingredient for overcoming challenges in robotic manipulation tasks with sparse rewards. A video showing experimental results is available at https://youtu.be/cLRrkd3Y7vU

📄 PDF Abstract BibTeX

Code (1)

misc-project/misc 공식 구현

Tasks

OpenAI Gymreinforcement-learningReinforcement LearningReinforcement Learning (RL)

Similar Papers 제목 키워드 기반

Intrinsically Motivated Self-supervised Learning in Reinforcement Learning

2021-06-26 · Yue Zhao, Chenzhuang Du, Hang Zhao, Tiejun Li

In vision-based reinforcement learning (RL) tasks, it is prevalent to assign auxiliary tasks with a surrogate self-supervised loss so as to obtain more semantic representations and improve sample efficiency. However, abu…

Decision Makingreinforcement-learningReinforcement LearningReinforcement Learning (RL)+2

Using Multiple Self-Supervised Tasks Improves Model Robustness

2022-04-07 · Matthew Lawhon, Chengzhi Mao, Junfeng Yang

Deep networks achieve state-of-the-art performance on computer vision tasks, yet they fail under adversarial attacks that are imperceptible to humans. In this paper, we propose a novel defense that can dynamically adapt …

Pretraining Neural Architecture Search Controllers with Locality-based Self-Supervised Learning

2021-03-15 · Kwanghee Choi, Minyoung Choe, Hyelee Lee

Neural architecture search (NAS) has fostered various fields of machine learning. Despite its prominent dedications, many have criticized the intrinsic limitations of high computational cost. We aim to ameliorate this by…

Metric LearningNeural Architecture SearchSelf-Supervised Learning

Mutual Information State Intrinsic Control

2021-03-15 · ICLR 2021 1 · Rui Zhao, Yang Gao, Pieter Abbeel, Volker Tresp 외

Reinforcement learning has been shown to be highly successful at many challenging tasks. However, success heavily relies on well-shaped rewards. Intrinsically motivated RL attempts to remove this constraint by defining a…

Physically Controllable Relighting of Photographs

2025-08-07 · Chris Careaga, Yağız Aksoy arxiv

We present a self-supervised approach to in-the-wild image relighting that enables fully controllable, physically based illumination editing. We achieve this by combining the physical accuracy of traditional rendering wi…

Image Relighting