Summary of A Large Recurrent Action Model: Xlstm Enables Fast Inference For Robotics Tasks, by Thomas Schmied et al.
A Large Recurrent Action Model: xLSTM enables Fast Inference for Robotics Tasks
by Thomas Schmied, Thomas Adler, Vihang Patil, Maximilian Beck, Korbinian Pöppel, Johannes Brandstetter, Günter Klambauer, Razvan Pascanu, Sepp Hochreiter
First submitted to arxiv on: 29 Oct 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper explores the use of modern recurrent architectures, specifically xLSTM and Mamba, for large action models in Reinforcement Learning (RL). The goal is to create agents that excel in both performance and inference time. To achieve this, the authors propose a Large Recurrent Action Model (LRAM) with an xLSTM core, which offers linear-time inference complexity and natural sequence length extrapolation abilities. The LRAM is tested on 432 tasks from six domains, showing comparable performance to Transformer-based models while being significantly faster. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper looks at how to make agents better for real-world uses. Right now, most RL agents are really good but take a long time to make decisions. To fix this, the authors try using new types of recurrent networks (like xLSTM and Mamba) that can work fast and well. They create a new model called LRAM that combines these new networks with an older one called xLSTM. Then, they test LRAM on lots of different tasks and show it does almost as well as the best current agents but way faster. |
Keywords
» Artificial intelligence » Inference » Reinforcement learning » Transformer
