Typesafe Abstractions for Tensor Operations (Short Paper)

被引：5

作者：

Chen, Tongfei ^{[1
]}

机构：

[1] Johns Hopkins Univ, Baltimore, MD 21218 USA

来源：

SCALA'17: PROCEEDINGS OF THE 8TH ACM SIGPLAN INTERNATIONAL SYMPOSIUM ON SCALA | 2017年

关键词：

Scala; tensor; heterogeneous list; deep learning;

D O I：

10.1145/3136000.3136001

中图分类号：

TP31 [计算机软件];

学科分类号：

081202 ; 0835 ;

摘要：

We propose a typesafe abstraction to tensors (i.e. multi-dimensional arrays) exploiting the type-level programming capabilities of Scala through heterogeneous lists (HList), and showcase typesafe abstractions of common tensor operations and various neural layers such as convolution or recurrent neural networks. This abstraction could lay the foundation of future typesafe deep learning frameworks that runs on Scala/JVM.

引用

页码：45 / 50

页数：6

共 17 条

[1] Abadi Martin, 2015, TENSOR FLOW LARGESCA
[2] [Anonymous], 2008, EVALUATING DERIVATIV
[3] Codd E. F., 1979, ACM Transactions on Database Systems, V4, P397, DOI 10.1145/320107.320109
[4] Eaton Frederik, 2006, P 2006 ACM SIGPLAN W, P120, DOI [10.1145/1159842.1159859, DOI 10.1145/1159842.1159859]
[5] Griffoen P. R., 2015, P 27 S IMPL APPL FUN, DOI 10.1145/2897336.2897341
[6] Hochreiter S, 1997, NEURAL COMPUT, V9, P1735, DOI [10.1162/neco.1997.9.1.1, 10.1007/978-3-642-24797-2]
[7] Generalized algebraic data types and object-oriented programming
Kennedy, A
Russo, CV
[J]. ACM SIGPLAN NOTICES, 2005, 40 (10) : 21 - 40
[8] Koller D, 2009, Probabilistic graphical models: principles and techniques
[9] Muranushi T, 2014, ACM SIGPLAN NOTICES, V49, P31, DOI [10.1145/2633357.2633362, 10.1145/2775050.2633362]
[10] Neubig Graham, 2017, DyNet: The Dynamic Neural Network Toolkit

← 1 2 →