Summary of Graph Is All You Need? Lightweight Data-agnostic Neural Architecture Search Without Training, by Zhenhan Huang et al.
Graph is all you need? Lightweight data-agnostic neural architecture search without training
by Zhenhan Huang, Tejaswini Pedapati, Pin-Yu Chen, Chunhen Jiang, Jianxi Gao
First submitted to arxiv on: 2 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
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 nasgraph method significantly reduces the computational costs required for training neural network models designed by Neural Architecture Search (NAS) algorithms. By converting neural architectures into graphs and using the average degree as a proxy evaluation metric, nasgraph eliminates the need for training candidate models, making it data-agnostic and lightweight. This approach enables finding the best architecture among 200 randomly sampled candidates from NAS-Bench201 in just 217 CPU seconds. The method achieves competitive performance on various datasets, including NASBench-101, NASBench-201, and NDS search spaces. Additionally, nasgraph generalizes to more challenging tasks on Micro TransNAS-Bench-101. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary nasgraph is a new way to design neural networks that doesn’t require training many different models. Instead, it converts the architectures into special graphs and uses a measure called average degree to decide which one works best. This makes it much faster than traditional methods and can even work with different types of data. nasgraph has been tested on several datasets and performs well, even when trying harder tasks. |
Keywords
» Artificial intelligence » Neural network
