Summary of Kan or Mlp: a Fairer Comparison, by Runpeng Yu and Weihao Yu and Xinchao Wang
KAN or MLP: A Fairer Comparison
by Runpeng Yu, Weihao Yu, Xinchao Wang
First submitted to arxiv on: 23 Jul 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 presents a comprehensive comparison of KAN and MLP models across various tasks, including machine learning, computer vision, audio processing, natural language processing, and symbolic formula representation. The study controls the number of parameters and FLOPs to compare the performance of KAN and MLP, finding that MLP generally outperforms KAN in most tasks. However, KAN excels in symbolic formula representation tasks, particularly when using its B-spline activation function. The paper also explores the forgetting issue in KAN, which is more severe than in MLP. This research aims to provide insights for future studies on KAN and other MLP alternatives. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper compares two types of artificial intelligence models, called KAN and MLP. It looks at how well these models do different tasks like recognizing images or understanding speech. The researchers made sure both models had the same number of “brain cells” (called parameters) and used the same amount of computer power (called FLOPs). They found that one model, MLP, usually did better than the other, KAN. But there was one task where KAN did really well because it has a special way of working called B-spline. The researchers also looked at how well these models remembered things they had learned before. They found that KAN forgot more easily than MLP. This study helps us understand which model is better for certain tasks and what makes them different. |
Keywords
* Artificial intelligence * Machine learning * Natural language processing
