mlsanityData to 7 Oct 2026

Compositional Multi-Object Reinforcement Learning with Linear Relation Networks

arXiv 2201.13388 ↗First posted 31 Jan 2022cs.RO cs.LG stat.MLcited by 19

Not on the map. Why?

Abstract (arXiv)

Although reinforcement learning has seen remarkable progress over the last years, solving robust dexterous object-manipulation tasks in multi-object settings remains a challenge. In this paper, we focus on models that can learn manipulation tasks in fixed multi-object settings and extrapolate this skill zero-shot without any drop in performance when the number of objects changes. We consider the generic task of bringing a specific cube out of a set to a goal position. We find that previous approaches, which primarily leverage attention and graph neural network-based architectures, do not generalize their skills when the number of input objects changes while scaling as K2. We propose an alternative plug-and-play module based on relational inductive biases to overcome these limitations. Besides exceeding performances in their training environment, we show that our approach, which scales linearly in K, allows agents to extrapolate and generalize zero-shot to any new object number.

Key facts. Cited by 19 papers in the map (data through 7 Oct 2026). Builds on 36 of its 56 references that are in the map. First posted on arXiv 31 Jan 2022.

Builds on

36 of 56 references.

References · 36

  1. 2017Attention Is All You Needcited by 59,932
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2017Proximal Policy Optimization Algorithmscited by 14,441
  4. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,282
  5. 2017Graph Attention Networkscited by 8,854
  6. 2016Layer Normalizationcited by 5,579
  7. 2021Decision Transformer: Reinforcement Learning via Sequence Modelingcited by 1,528
  8. 2018Relational inductive biases, deep learning, and graph networkscited by 1,341
  9. 2017Learning to Compare: Relation Network for Few-Shot Learningcited by 1,139
  10. 2021Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gaugescited by 1,131
  11. 2020Object-Centric Learning with Slot Attentioncited by 865
  12. 2017Learning Complex Dexterous Manipulation with Deep Reinforcement Learning and Demonstrationscited by 685
24 more
  1. 2018Learning Dexterous In-Hand Manipulationcited by 643
  2. 2016Interaction Networks for Learning about Objects, Relations and Physicscited by 503
  3. 2019Multi-Object Representation Learning with Iterative Variational Inferencecited by 359
  4. 2020Curriculum Learning for Reinforcement Learning Domains: A Framework and Surveycited by 350
  5. 2017A simple neural network module for relational reasoningcited by 345
  6. 2017Relation Networks for Object Detectioncited by 305
  7. 2021Conditional Object-Centric Learning from Videocited by 247
  8. 2017One-Shot Imitation Learningcited by 232
  9. 2020On the Binding Problem in Artificial Neural Networkscited by 230
  10. 2020CausalWorld: A Robotic Manipulation Benchmark for Causal Structure and Transfer Learningcited by 109
  11. 2018Relational Deep Reinforcement Learningcited by 72
  12. 2021Beyond Pick-and-Place: Tackling Robotic Stacking of Diverse Shapescited by 70
  13. 2019Towards Practical Multi-Object Manipulation using Relational Reinforcement Learningcited by 54
  14. 2018Reasoning About Physical Interactions with Object-Oriented Prediction and Planningcited by 51
  15. 2020Self-supervised Visual Reinforcement Learning with Object-centric Representationscited by 43
  16. 2016Towards Deep Symbolic Reinforcement Learningcited by 42
  17. 2019rlpyt: A Research Code Base for Deep Reinforcement Learning in PyTorchcited by 38
  18. 2021Transferring Dexterous Manipulation from GPU Simulation to a Remote Real-World TriFingercited by 34
  19. 2020TriFinger: An Open-Source Robot for Learning Dexteritycited by 31
  20. 2019On Inductive Biases in Deep Reinforcement Learningcited by 22
  21. 2021Benchmarking Structured Policies and Policy Optimization for Real-World Dexterous Object Manipulationcited by 14
  22. 2021The Role of Pretrained Representations for the OOD Generalization of Reinforcement Learning Agentscited by 12
  23. 2017Discovering objects and their relations from entangled scene representationscited by 11
  24. 2019Learning to Manipulate Object Collections Using Grounded State Representationscited by 11

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 3; 2023: 9; 2024: 3; 2025: 1; 2026: 3.

19 papers.

Sort:

By area:

  1. 2022Bridging the Gap to Real-World Object-Centric Learningcited by 219
  2. 2022Simple Unsupervised Object-Centric Learning for Complex and Naturalistic Videoscited by 139
  3. 2023Object-Centric Learning for Real-World Videos by Predicting Temporal Feature Similaritiescited by 78
  4. 2023An Investigation into Pre-Training Object-Centric Representations for Reinforcement Learningcited by 49
  5. 2023OCAtari: Object-Centric Atari 2600 Reinforcement Learning Environmentscited by 30
  6. 2024Entity-Centric Reinforcement Learning for Object Manipulation from Pixelscited by 27
  7. 2023Slot-VAE: Object-Centric Scene Generation with Slot Attentioncited by 20
  8. 2023Structure in Deep Reinforcement Learning: A Survey and Open Problemscited by 16
  9. 2023Learning Dynamic Attribute-factored World Models for Efficient Multi-object Reinforcement Learningcited by 15
  10. 2023Object-Centric Multiple Object Trackingcited by 15
  11. 2022Inductive Biases for Object-Centric Representations in the Presence of Complex Texturescited by 13
  12. 2023A Survey on Compositional Generalization in Applicationscited by 12
7 more
  1. 2024Entity-based Reinforcement Learning for Autonomous Cyber Defencecited by 11
  2. 2025Learning Interactive World Model for Object-Centric Reinforcement Learningcited by 11
  3. 2023Imagine the Unseen World: A Benchmark for Systematic Generalization in Visual World Modelscited by 6
  4. 2026Hierarchical Entity-centric Reinforcement Learning with Factored Subgoal Diffusioncited by 6
  5. 2024Object-Centric Temporal Consistency via Conditional Autoregressive Inductive Biasescited by 5
  6. 2026Slot-MPC: Goal-Conditioned Model Predictive Control with Object-Centric Representationscited by 3
  7. 2026QASA: Quality-Guided K-Adaptive Slot Attention for Unsupervised Object-Centric Learningcited by 0