Hypervector Design for Efficient Hyperdimensional Computing on Edge Devices
Hypervector Design for Efficient Hyperdimensional Computing on Edge Devices
复制标题
DOI：
--
复制
发表时间：
2021-03
期刊：
影响因子：
--
通讯作者：
T. Basaklar;Y. Tuncel;S. Y. Narayana;S. Gumussoy;U. Ogras
中科院分区：
文献类型：
--
作者：
T. Basaklar;Y. Tuncel;S. Y. Narayana;S. Gumussoy;U. Ogras
超维计算（HDC）作为一种新的轻量级学习算法，与传统技术相比具有更小的计算量和能量需求。在HDC中，数据点由高维向量（hypervector）表示，并映射到高维空间（hyperspace）。通常，需要一个大的超向量维度（$\geq1000$）来实现与传统替代方案相当的精度。然而，不必要的大型超向量会增加硬件和能源成本，这可能会破坏它们的好处。本文提出了一种最小化超向量维数的方法，同时保持分类器的准确性和提高分类器的鲁棒性。为此，我们在文献中首次将超向量设计表述为一个多目标优化问题。该方法在保持或提高常规HDC精度的同时，将超向量维数降低了$32\times$以上。在商业硬件平台上的实验表明，该方法在模型尺寸、推理时间和能耗方面都降低了一个数量级以上。我们还展示了精度和对噪声的鲁棒性之间的权衡，并提供了帕累托前解作为超矢量设计中的设计参数。
Hyperdimensional computing (HDC) has emerged as a new light-weight learning algorithm with smaller computation and energy requirements compared to conventional techniques. In HDC, data points are represented by high-dimensional vectors (hypervectors), which are mapped to high-dimensional space (hyperspace). Typically, a large hypervector dimension ($\geq1000$) is required to achieve accuracies comparable to conventional alternatives. However, unnecessarily large hypervectors increase hardware and energy costs, which can undermine their benefits. This paper presents a technique to minimize the hypervector dimension while maintaining the accuracy and improving the robustness of the classifier. To this end, we formulate the hypervector design as a multi-objective optimization problem for the first time in the literature. The proposed approach decreases the hypervector dimension by more than $32\times$ while maintaining or increasing the accuracy achieved by conventional HDC. Experiments on a commercial hardware platform show that the proposed approach achieves more than one order of magnitude reduction in model size, inference time, and energy consumption. We also demonstrate the trade-off between accuracy and robustness to noise and provide Pareto front solutions as a design parameter in our hypervector design.