Summary of Gaussian Process Neural Additive Models, by Wei Zhang et al.
Gaussian Process Neural Additive Models
by Wei Zhang, Brian Barr, John Paisley
First submitted to arxiv on: 19 Feb 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 proposed Gaussian Process Neural Additive Models (GP-NAM) is a subclass of Neural Additive Models (NAMs) designed for interpretable deep learning on tabular datasets. By using random Fourier features to construct single-layer neural networks, GP-NAM achieves a convex objective function and linearly growing trainable parameters with feature dimensionality. This approach suffers no performance loss compared to deeper NAMs due to the well-suited Gaussian Process (GP) for learning complex non-parametric univariate functions. The model’s effectiveness is demonstrated on several tabular datasets, showcasing comparable or better performance in both classification and regression tasks while reducing the number of parameters. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper proposes a new type of artificial intelligence called GP-NAM that can help make deep neural networks more understandable and trustworthy. This is important because some fields like healthcare and finance need models they can explain. The researchers use something called random Fourier features to build a simple layer of neural networks that can learn complex patterns in data. This approach is better than others because it’s easier to understand and train, and it works just as well on different types of data. |
Keywords
* Artificial intelligence * Classification * Deep learning * Objective function * Regression
