Summary of On the Diminishing Returns Of Width For Continual Learning, by Etash Guha et al.
On the Diminishing Returns of Width for Continual Learning
by Etash Guha, Vihan Lakshman
First submitted to arxiv on: 11 Mar 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 The abstract proposes a novel framework for analyzing Continual Learning Theory, focusing on the relationship between neural network width and catastrophic forgetting. The authors prove that increasing width leads to reduced forgetting in Feed-Forward Networks (FFN), but also show that this approach yields diminishing returns. Empirical evidence is provided to support these claims, demonstrating unexplored widths where predicted diminishing returns are observed. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper explores how neural networks can learn new tasks without forgetting previous ones. Researchers have found that making the network wider helps reduce forgetting, but they didn’t know exactly how this works. The authors created a framework to study this and showed that increasing width does help, but only up to a point. After that, it doesn’t make as much difference. This is important because it can help us design better networks for learning new things. |
Keywords
* Artificial intelligence * Continual learning * Neural network
