Summary of Gats: Gather-attend-scatter, by Konrad Zolna et al.
GATS: Gather-Attend-Scatter
by Konrad Zolna, Serkan Cabi, Yutian Chen, Eric Lau, Claudio Fantacci, Jurgis Pasukonis, Jost Tobias Springenberg, Sergio Gomez Colmenarejo
First submitted to arxiv on: 16 Jan 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Robotics (cs.RO)
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 introduces Gather-Attend-Scatter (GATS), a novel module that enables seamless integration of large-scale foundation models into larger multimodal networks. GATS allows AI systems to process and generate information across multiple modalities at different rates, without the risk of component models losing important knowledge acquired during pretraining. The authors demonstrate GATS’ utility and versatility in experiments involving games, robotics, and multimodal input-output systems. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary GATS is a new way to combine big AI models into even bigger ones. This helps computers understand and create information from many different sources at once. It’s like having a team of experts working together to solve problems. The authors show how GATS can be used in games, robots, and other systems that use multiple types of data. |
Keywords
* Artificial intelligence * Pretraining
