Summary of Neural Architecture Search Of Hybrid Models For Npu-cim Heterogeneous Ar/vr Devices, by Yiwei Zhao et al.
Neural Architecture Search of Hybrid Models for NPU-CIM Heterogeneous AR/VR Devices
by Yiwei Zhao, Ziyun Li, Win-San Khwa, Xiaoyu Sun, Sai Qian Zhang, Syed Shakib Sarwar, Kleber Hugo Stangherlin, Yi-Lun Lu, Jorge Tomas Gomez, Jae-Sun Seo, Phillip B. Gibbons, Barbara De Salvo, Chiao Liu
First submitted to arxiv on: 10 Oct 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Hardware Architecture (cs.AR); Machine Learning (cs.LG); Performance (cs.PF)
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 paper introduces a novel Neural Architecture Search (NAS) framework, H4H-NAS, which efficiently designs hybrid convolutional neural networks (CNNs) and vision transformers (ViTs) for edge AI applications. The framework leverages the architecture heterogeneity of Neural Processing Units (NPUs) and Compute-In-Memory (CIM) to execute these hybrid models, achieving superior accuracy and performance tradeoff on computer vision tasks. The authors also demonstrate significant improvements in top-1 accuracy (up to 1.34%) and latency/energy efficiency (up to 56.08% and 41.72%, respectively) compared to baseline solutions. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper helps create better virtual reality and augmented reality experiences by making AI work more efficiently on devices like smartphones or tablets. It uses a new way of designing artificial intelligence models that combines the strengths of different types of neural networks, called CNNs and ViTs. This approach allows for faster and more accurate processing of images, which is important for applications like object detection and image classification. |
Keywords
» Artificial intelligence » Image classification » Object detection
