methodology-9

Multi-Goal Reinforcement Learning

3260 papers • 126 benchmarks • 313 datasets

This task has no description! Would you like to contribute one?

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in multi-goal-reinforcement-learning-9

Trend

Dataset

Best Model

Actions

No benchmarks available.

Libraries

i

Use these libraries to find multi-goal-reinforcement-learning-9 models and implementations

Datasets

UniMiB SHAR

bipedal-skills

Subtasks

No subtasks available.

Most implemented papers

Multi-Goal Reinforcement Learning: Challenging Robotics Environments and Request for Research

Jonas Schneider, Vikash Kumar, Wojciech Zaremba, Matthias Plappert, Bowen Baker, Alex Ray, Peter Welinder, Bob McGrew, Marcin Andrychowicz, Glenn Powell, Joshua Tobin, Maciek Chociej•Sun Feb 25 2018

A suite of challenging continuous control tasks (integrated with OpenAI Gym) based on currently existing robotics hardware and following a Multi-Goal Reinforcement Learning (RL) framework are introduced.

620

Content

0

Paper Graph

Intrinsically Motivated Goal Exploration Processes with Automatic Curriculum Learning

Pierre-Yves Oudeyer, Sébastien Forestier, Yoan Mollard•Sun Aug 06 2017

It is illustrated the computational efficiency of IMGEPs as these robotic experiments use a simple memory-based low-level policy representations and search algorithm, enabling the whole system to learn online and incrementally on a Raspberry Pi 3.

199 0

Paper Graph

Maximum Entropy-Regularized Multi-Goal Reinforcement Learning

Volker Tresp, Rui Zhao, Xudong Sun•Mon May 20 2019

A novel multi-goal RL objective based on weighted entropy is proposed, which encourages the agent to maximize the expected return, as well as to achieve more diverse goals and a maximum entropy-based prioritization framework is developed to optimize the proposed objective.

92 0

Paper Graph

Learning to Reach Goals via Iterated Supervised Learning

S. Levine, Justin Fu, Abhishek Gupta, Benjamin Eysenbach, Coline Devin, Dibya Ghosh, Ashwin Reddy•Wed Dec 11 2019

This paper proposes a simple algorithm in which an agent continually relabels and imitates the trajectories it generates to progressively learn goal- reaching behaviors from scratch, and formally shows that this iterated supervised learning procedure optimizes a bound on the RL objective, derive performance bounds of the learned policy, and empirically demonstrates improved goal-reaching performance and robustness over current RL algorithms in several benchmark tasks.

208 0

Paper Graph

An Inductive Bias for Distances: Neural Nets that Respect the Triangle Inequality

Jimmy Ba, Silviu Pitis, Harris Chan, Kiarash Jamali•Thu Feb 13 2020

Novel architectures that are guaranteed to satisfy the triangle inequality are introduced and it is shown that these architectures outperform existing metric approaches when modeling graph distances and have a better inductive bias than non-metric approaches when training data is limited in the multi-goal reinforcement learning setting.

30 0

Paper Graph

Maximum Entropy Gain Exploration for Long Horizon Multi-goal Reinforcement Learning

Jimmy Ba, Silviu Pitis, Harris Chan, S. Zhao, Bradly C. Stadie•Sun Jul 05 2020

This paper proposes to optimize this objective by having the agent pursue past achieved goals in sparsely explored areas of the goal space, which focuses exploration on the frontier of the achievable goal set.

141 0

Paper Graph

An Open-Source Multi-Goal Reinforcement Learning Environment for Robotic Manipulation with Pybullet

Yu-Kun Lai, Xintong Yang, Ze Ji, Jing Wu•Tue May 11 2021

This work re-implements the OpenAI Gym multi-goal robotic manipulation environment, originally based on the commercial Mujoco engine, onto the open-source Pybullet engine. By comparing the performances of the Hindsight Experience Replay-aided Deep Deterministic Policy Gradient agent on both environments, we demonstrate our successful re-implementation of the original environment. Besides, we provide users with new APIs to access a joint control mode, image observations and goals with customisable camera and a built-in on-hand camera. We further design a set of multi-step, multi-goal, long-horizon and sparse reward robotic manipulation tasks, aiming to inspire new goal-conditioned reinforcement learning algorithms for such challenges. We use a simple, human-prior-based curriculum learning method to benchmark the multi-step manipulation tasks. Discussions about future research opportunities regarding this kind of tasks are also provided.

21 0

Paper Graph

Grounding Hindsight Instructions in Multi-Goal Reinforcement Learning for Robotics

Stefan Wermter, Frank Röder, Manfred Eppe•Thu Apr 07 2022

This paper presents a mechanism for hindsight instruction replay utilizing expert feedback, a seq2seq model to generate linguistic hindsight instructions, and presents a novel class of language-focused learning tasks.

7 0

Paper Graph

CURIOUS: Intrinsically Motivated Modular Multi-Goal Reinforcement Learning

Pierre-Yves Oudeyer, Cédric Colas, Olivier Sigaud, Pierre Fournier, M. Chetouani•Sun Oct 14 2018

CURIOUS is proposed, an algorithm that leverages a modular Universal Value Function Approximator with hindsight learning to achieve a diversity of goals of different kinds within a unique policy and an automated curriculum learning mechanism that biases the attention of the agent towards goals maximizing the absolute learning progress.

182 0

Paper Graph

Bias-Reduced Hindsight Experience Replay with Virtual Goal Prioritization

A. Biess, Binyamin Manela•Mon May 13 2019

This paper presents two improvements over the existing HER algorithm, which prioritize virtual goals from which the agent will learn more valuable information, and reduces existing bias in HER by the removal of misleading samples.

24 0

Paper Graph

Adding a benchmark result helps the community track progress.