Neural NID Rules
Not on the map. Why?
Abstract (arXiv)
Abstract object properties and their relations are deeply rooted in human common sense, allowing people to predict the dynamics of the world even in situations that are novel but governed by familiar laws of physics. Standard machine learning models in model-based reinforcement learning are inadequate to generalize in this way. Inspired by the classic framework of noisy indeterministic deictic (NID) rules, we introduce here Neural NID, a method that learns abstract object properties and relations between objects with a suitably regularized graph neural network. We validate the greater generalization capability of Neural NID on simple benchmarks specifically designed to assess the transition dynamics learned by the model.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 11 of its 20 references that are in the map. First posted on arXiv 12 Feb 2022.
Builds on
11 of 20 references.
References · 11
- 2018Relational inductive biases, deep learning, and graph networkscited by 1,341
- 2020Object-Centric Learning with Slot Attentioncited by 865
- 2016Interaction Networks for Learning about Objects, Relations and Physicscited by 503
- 2020Discovering Symbolic Models from Deep Learning with Inductive Biasescited by 335
- 2019Contrastive Learning of Structured World Modelscited by 206
- 2017Neural Expectation Maximizationcited by 123
- 2019Entity Abstraction in Visual Model-Based Reinforcement Learningcited by 110
- 2019Unsupervised Learning of Object Keypoints for Perception and Controlcited by 109
- 2018Relational Neural Expectation Maximization: Unsupervised Discovery of Objects and their Interactionscited by 100
- 2019Neural Logic Reinforcement Learningcited by 53
- 2020PDDLGym: Gym Environments from PDDL Problemscited by 45
Built on by
Not cited yet.