Statheros: Compiler for Efficient Low-Precision Probabilistic Programming
Statheros: Compiler for Efficient Low-Precision Probabilistic Programming
复制标题
DOI：
10.1109/dac18074.2021.9586276
复制
发表时间：
2021-12
期刊：
影响因子：
--
通讯作者：
Jacob S. Laurel;Rem Yang;Atharva Sehgal;Shubham Ugare;Sasa Misailovic
中科院分区：
文献类型：
--
作者：
Jacob S. Laurel;Rem Yang;Atharva Sehgal;Shubham Ugare;Sasa Misailovic
由于边缘和物联网计算设备处理噪声数据或在不确定的环境中做出决策，它们需要廉价但准确的概率推理框架。概率编程已成为开发人员编写高级程序的一种强大方式，同时抽象出推理的实现细节。然而，现有的算法是缓慢的，往往被认为需要精确的计算。我们提出了Statheros，第一个编译器的低层次，不动点近似的概率编程。Statheros将程序编译为定点推理过程，并能够确定要使用的最佳定点类型。我们在13个基准测试和三个嵌入式平台上评估Statheros。结果显示，Statheros生成的代码为11。5x（Arduino），3. 8x（PocketBeagle）和2.比单精度浮点计算快2倍（Raspberry Pi），精度损失最小。
As Edge and IoT computing devices process noisy data or make decisions in uncertain environments, they require frameworks for inexpensive, yet accurate probabilistic inference. Probabilistic programming has emerged as a powerful way for developers to write high-level programs, while abstracting away the implementation details of inference. However, the existing algorithms are slow and often assumed to require precise calculations. We present Statheros, the first compiler for low-level, fixed-point approximation of probabilistic programming. Statheros compiles programs to fixed-point inference procedures and is able to determine the optimal fixed-point type to use. We evaluate Statheros on 13 benchmarks and three embedded platforms. The results show that Statheros-generated code is 11. 5x (Arduino), 3. 8x (PocketBeagle), and 2. 2x (Raspberry Pi) faster than single-precision floating-point computation, with minimal accuracy loss.