Decision Mamba Architectures

André Correia, Luís A. Alexandre
5/13/2024
cs.LGcs.AI

Official code: meowatthemoon/DecisionMamba (1 star, last commit 2024-07-23). 1 more implementation below.

Abstract

Recent advancements in imitation learning have been largely fueled by the integration of sequence models, which provide a structured flow of information to effectively mimic task behaviours. Currently, Decision Transformer (DT) and subsequently, the Hierarchical Decision Transformer (HDT), presented Transformer-based approaches to learn task policies. Recently, the Mamba architecture has shown to outperform Transformers across various task domains. In this work, we introduce two novel methods, Decision Mamba (DM) and Hierarchical Decision Mamba (HDM), aimed at enhancing the performance of the Transformer models. Through extensive experimentation across diverse environments such as OpenAI Gym and D4RL, leveraging varying demonstration data sets, we demonstrate the superiority of Mamba models over their Transformer counterparts in a majority of tasks. Results show that DM outperforms other methods in most settings. The code can be found at https://github.com/meowatthemoon/DecisionMamba.

View on arXivView PDF

Code Implementations(2)

meowatthemoon/DecisionMambaOfficial100%

Official Implementation of the Decision Mamba model from the Decision Mamba Architectures paper.

10PythonMar 29, 20242 years ago
meowatthemoon/HierarchicalDecisionMambaCommunity100%

Official Implementation of the Hierarchical Decision Mamba model from the Decision Mamba Architectures paper.

41PythonMar 29, 20242 years ago

Similar Papers

Decision Mamba: Reinforcement Learning via Sequence Modeling with Selective State Spaces

88% match1 repo542024

Decision Mamba: Reinforcement Learning via Hybrid Selective Sequence Modeling

87% match1 repo542024

Integrating Multi-Modal Input Token Mixer Into Mamba-Based Decision Models: Decision MetaMamba

86% match2024

MaIL: Improving Imitation Learning with Mamba

83% match1 repo482024

Is Mamba Compatible with Trajectory Optimization in Offline Reinforcement Learning?

83% match1 repo52024

Decision Transformer vs. Decision Mamba: Analysing the Complexity of Sequential Decision Making in Atari Games

82% match1 repo22024

Cite this paper

@article{correia2024decision,
  title  = {Decision Mamba Architectures},
  author = {André Correia and Luís A. Alexandre},
  year   = {2024},
  eprint = {2405.07943},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2405.07943v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2405.07943.svg)](https://codeofpaper.com/paper/2405.07943)

Discussion