Summary of Multi-objective Hardware Aware Neural Architecture Search Using Hardware Cost Diversity, by Nilotpal Sinha et al.
Multi-Objective Hardware Aware Neural Architecture Search using Hardware Cost Diversity
by Nilotpal Sinha, Peyman Rostami, Abd El Rahman Shabayek, Anis Kacem, Djamila Aouada
First submitted to arxiv on: 15 Apr 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 A novel approach to Hardware-aware Neural Architecture Search (HW-NAS) has been developed, called MO-HDNAS, which tackles the issue of expensive architecture performance evaluation by proposing a Multi-Objective method. This method optimizes three objectives simultaneously: maximizing representation similarity metric, minimizing hardware cost, and maximizing hardware cost diversity. By doing so, it allows for a better exploration of the architecture search space and reduces computational costs. Experimental results demonstrate the effectiveness of MO-HDNAS in efficiently addressing HW-NAS problems across six edge devices for image classification tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine you’re trying to design a new house, but instead of using traditional blueprints, you’re working with tiny LEGO blocks that can be arranged in different ways to create different rooms. This is similar to the problem faced by computer scientists when designing special kinds of computers called neural networks. To make this process easier and more efficient, researchers have developed a new method called MO-HDNAS. This method allows them to design multiple neural networks at once that work well on different types of devices, such as smartphones or tablets. |
Keywords
» Artificial intelligence » Image classification
