Intelligent Systems
Note: This research group has relocated.

Neuro-algorithmic Policies Enable Fast Combinatorial Generalization

2021

Conference Paper

al


Although model-based and model-free approa\-ches to learning the control of systems have achieved impressive results on standard benchmarks, generalization to task variations is still lacking. Recent results suggest that generalization for standard architectures improves only after obtaining exhaustive amounts of data. We give evidence that generalization capabilities are in many cases bottlenecked by the inability to generalize on the combinatorial aspects of the problem. We show that, for a certain subclass of the MDP framework, this can be alleviated by a neuro-algorithmic policy architecture that embeds a time-dependent shortest path solver in a deep neural network. Trained end-to-end via blackbox-differentiation, this method leads to considerable improvement in generalization capabilities in the low-data regime.

Author(s): Marin Vlastelica and Michal Rolinek and Georg Martius
Book Title: Proceedings of the 2021 International Conference on Machine Learning (ICML)
Year: 2021
Month: July

Department(s): Autonomous Learning
Research Project(s): Combinatorial Optimization as a Layer / Blackbox Differentiation
Bibtex Type: Conference Paper (inproceedings)
Paper Type: Conference

Event Name: The Thirty-eighth International Conference on Machine Learning (ICML)
Event Place: Virtual

Links: arXiv
Spotlight
PDF

BibTex

@inproceedings{VlastelicaEtal2021:NeuroAlgorithmic,
  title = {Neuro-algorithmic Policies Enable Fast Combinatorial Generalization},
  author = {Vlastelica, Marin and Rolinek, Michal and Martius, Georg},
  booktitle = {Proceedings of the 2021 International Conference on Machine Learning (ICML)},
  month = jul,
  year = {2021},
  doi = {},
  month_numeric = {7}
}