Boosted Race Trees for Low Energy Classification

Boosted Race Trees for Low Energy Classification
复制标题

用于低能量分类的增强竞赛树

DOI:
10.1145/3297858.3304036
复制
发表时间:
2019
期刊:
Proceedings of the Twenty-Fourth International Conference on Architectural Support for Programming Languages and Operating Systems Pages
影响因子:
--
通讯作者:
Sherwood, Timothy
Sherwood, Timothy
中科院分区:
--
文献类型:
--
作者:
Tzimpragos, Georgios;Madhavan, Advait;Vasudevan, Dilip;Strukov, Dmitri;Sherwood, Timothy

文献摘要

参考文献

被引文献

相似文献

当需要极低能耗的处理时,数据表示的选择会产生巨大的差异。每种表示(例如,频域、残差编码、对数尺度)都有一组独特的权衡-一些运算在那个域中更容易,而另一些则更难。我们展示了RACE逻辑,其中时间编码的信号以数据流的方式被处理,为传感器内处理应用提供了有趣的新功能。具体地说,通过一组扩展的竞争逻辑运算,我们证明了基于树的分类器可以被自然地编码,并且常见的分类任务可以作为这类逻辑中的可编程加速器有效地实现。为了验证这一假设,我们设计了几个集成学习者的种族逻辑实现,将它们与最先进的分类器进行比较,并进行了建筑设计空间的探索。我们的概念验证架构由1,000个深度为6的可重构Race树组成,将处理15.2M帧/S,在14 nm CMOS中消耗613 mW。
When extremely low-energy processing is required, the choice of data representation makes a tremendous difference. Each representation (e.g. frequency domain, residue coded, log-scale) comes with a unique set of trade-offs --- some operations are easier in that domain while others are harder. We demonstrate that race logic, in which temporally coded signals are getting processed in a dataflow fashion, provides interesting new capabilities for in-sensor processing applications. Specifically, with an extended set of race logic operations, we show that tree-based classifiers can be naturally encoded, and that common classification tasks can be implemented efficiently as a programmable accelerator in this class of logic. To verify this hypothesis, we design several race logic implementations of ensemble learners, compare them against state-of-the-art classifiers, and conduct an architectural design space exploration. Our proof-of-concept architecture, consisting of 1,000 reconfigurable Race Trees of depth 6, will process 15.2M frames/s, dissipating 613mW in 14nm CMOS.
互连结构的未来:逆向观点
DOI: --
发表时间: 2010
期刊: International Workshop on System-Level Interconnect Prediction
影响因子: --
作者:
S. Borkar
通讯作者: S. Borkar
使用跨层仿真框架对在多个电压范围下运行的基于多阈值 7nm-FinFET 的电路进行性能预测
DOI: --
发表时间: 2014
期刊: IEEE SOI-3D-Subthreshold Microelectronics Technology Unified Conference
影响因子: --
作者:
Shuang Chen;Yanzhi Wang;Xue Lin;Q. Xie;Massoud Pedram
通讯作者: Massoud Pedram
可解释的人工智能
DOI: 10.29242/rli.299.3
发表时间: 2019
期刊: Research Handbook on Big Data Law
影响因子: --
作者:
M. Ridley
通讯作者: M. Ridley
首次尖峰时间 CMOS 成像仪
DOI: 10.1109/iscas.2004.1329131
发表时间: 2004
期刊: 2004 IEEE International Symposium on Circuits and Systems (IEEE Cat. No.04CH37512)
影响因子: --
作者:
Xin Qi;Xiaochuan Guo;J. Harris
通讯作者: J. Harris
Charm:一种封闭式高级架构建模语言
DOI: 10.1109/isca.2018.00023
发表时间: 2018
期刊: 2018 ACM/IEEE 45th Annual International Symposium on Computer Architecture (ISCA
影响因子: --
作者:
Cui, Weilong;Ding, Yongshan;Dangwal, Deeksha;Holmes, Adam;McMahan, Joseph;Javadi-Abhari, Ali;Tzimpragos, Georgios;Chong, Frederic;Sherwood, Timothy
通讯作者: Sherwood, Timothy