{
  "id": "1911.00105",
  "version": "v1",
  "published": "2019-10-31T21:02:23.000Z",
  "updated": "2019-10-31T21:02:23.000Z",
  "title": "On Neural Architecture Search for Resource-Constrained Hardware Platforms",
  "authors": [
    "Qing Lu",
    "Weiwen Jiang",
    "Xiaowei Xu",
    "Yiyu Shi",
    "Jingtong Hu"
  ],
  "comment": "8 pages, ICCAD 2019",
  "categories": [
    "cs.LG",
    "cs.NE",
    "eess.SP"
  ],
  "abstract": "In the recent past, the success of Neural Architecture Search (NAS) has enabled researchers to broadly explore the design space using learning-based methods. Apart from finding better neural network architectures, the idea of automation has also inspired to improve their implementations on hardware. While some practices of hardware machine-learning automation have achieved remarkable performance, the traditional design concept is still followed: a network architecture is first structured with excellent test accuracy, and then compressed and optimized to fit into a target platform. Such a design flow will easily lead to inferior local-optimal solutions. To address this problem, we propose a new framework to jointly explore the space of neural architecture, hardware implementation, and quantization. Our objective is to find a quantized architecture with the highest accuracy that is implementable on given hardware specifications. We employ FPGAs to implement and test our designs with limited loop-up tables (LUTs) and required throughput. Compared to the separate design/searching methods, our framework has demonstrated much better performance under strict specifications and generated designs of higher accuracy by 18\\% to 68\\% in the task of classifying CIFAR10 images. With 30,000 LUTs, a light-weight design is found to achieve 82.98\\% accuracy and 1293 images/second throughput, compared to which, under the same constraints, the traditional method even fails to find a valid solution.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2019-10-31T21:02:23.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "neural architecture search",
      "resource-constrained hardware platforms",
      "finding better neural network architectures",
      "traditional design concept",
      "inferior local-optimal solutions"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 8,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}