Timezone: »
Compositional generalization is a critical ability in learning and decision-making. We focus on the setting of reinforcement learning in object-oriented environments to study compositional generalization in world modeling. We (1) formalize the compositional generalization problem with an algebraic approach and (2) study how a world model can achieve that. We introduce a conceptual environment, Object Library, and two instances, and deploy a principled pipeline to measure the generalization ability. Motivated by the formulation, we analyze several methods with exact or no compositional generalization ability using our framework, and design a differentiable approach, Homomorphic Object-oriented World Model (HOWM), that achieves soft but more efficient compositional generalization.
Author Information
Linfeng Zhao (Northeastern University)
Lingzhi Kong (Northeastern University)
Robin Walters (Northeastern University)
Lawson Wong (Northeastern University)
Related Events (a corresponding poster, oral, or spotlight)
-
2022 Oral: Toward Compositional Generalization in Object-Oriented World Modeling »
Thu. Jul 21st 06:05 -- 06:25 PM Room Hall F
More from the Same Authors
-
2021 : Bad-Policy Density: A Measure of Reinforcement-Learning Hardness »
David Abel · Cameron Allen · Dilip Arumugam · D Ellis Hershkowitz · Michael L. Littman · Lawson Wong -
2023 : Unsupervised Learning of 3-colorings using Simplicial Higher-Order Neural Networks »
Lucas Laird · Robin Walters · Wolfgang Gatterbauer -
2023 : Latent Space Symmetry Discovery »
Jianke Yang · Nima Dehmamy · Robin Walters · Rose Yu -
2023 : Can Euclidean Symmetry Help in Reinforcement Learning and Planning »
Linfeng Zhao · Owen Howell · Jung Yeon Park · Xupeng Zhu · Robin Walters · Lawson Wong -
2023 Poster: Generative Adversarial Symmetry Discovery »
Jianke Yang · Robin Walters · Nima Dehmamy · Rose Yu -
2022 Poster: Learning Symmetric Embeddings for Equivariant World Models »
Jung Yeon Park · Ondrej Biza · Linfeng Zhao · Jan-Willem van de Meent · Robin Walters -
2022 Spotlight: Learning Symmetric Embeddings for Equivariant World Models »
Jung Yeon Park · Ondrej Biza · Linfeng Zhao · Jan-Willem van de Meent · Robin Walters -
2022 Poster: Approximately Equivariant Networks for Imperfectly Symmetric Dynamics »
Rui Wang · Robin Walters · Rose Yu -
2022 Spotlight: Approximately Equivariant Networks for Imperfectly Symmetric Dynamics »
Rui Wang · Robin Walters · Rose Yu -
2021 : Bad-Policy Density: A Measure of Reinforcement-Learning Hardness »
David Abel · Cameron Allen · Dilip Arumugam · D Ellis Hershkowitz · Michael L. Littman · Lawson Wong