-
Automatic differentiation in ML: Where we are and where we should be going
Authors:
Bart van Merriënboer,
Olivier Breuleux,
Arnaud Bergeron,
Pascal Lamblin
Abstract:
We review the current state of automatic differentiation (AD) for array programming in machine learning (ML), including the different approaches such as operator overloading (OO) and source transformation (ST) used for AD, graph-based intermediate representations for programs, and source languages. Based on these insights, we introduce a new graph-based intermediate representation (IR) which speci…
▽ More
We review the current state of automatic differentiation (AD) for array programming in machine learning (ML), including the different approaches such as operator overloading (OO) and source transformation (ST) used for AD, graph-based intermediate representations for programs, and source languages. Based on these insights, we introduce a new graph-based intermediate representation (IR) which specifically aims to efficiently support fully-general AD for array programming. Unlike existing dataflow programming representations in ML frameworks, our IR naturally supports function calls, higher-order functions and recursion, making ML models easier to implement. The ability to represent closures allows us to perform AD using ST without a tape, making the resulting derivative (adjoint) program amenable to ahead-of-time optimization using tools from functional language compilers, and enabling higher-order derivatives. Lastly, we introduce a proof of concept compiler toolchain called Myia which uses a subset of Python as a front end.
△ Less
Submitted 2 January, 2019; v1 submitted 26 October, 2018;
originally announced October 2018.
-
Theano: A Python framework for fast computation of mathematical expressions
Authors:
The Theano Development Team,
Rami Al-Rfou,
Guillaume Alain,
Amjad Almahairi,
Christof Angermueller,
Dzmitry Bahdanau,
Nicolas Ballas,
Frédéric Bastien,
Justin Bayer,
Anatoly Belikov,
Alexander Belopolsky,
Yoshua Bengio,
Arnaud Bergeron,
James Bergstra,
Valentin Bisson,
Josh Bleecher Snyder,
Nicolas Bouchard,
Nicolas Boulanger-Lewandowski,
Xavier Bouthillier,
Alexandre de Brébisson,
Olivier Breuleux,
Pierre-Luc Carrier,
Kyunghyun Cho,
Jan Chorowski,
Paul Christiano
, et al. (88 additional authors not shown)
Abstract:
Theano is a Python library that allows to define, optimize, and evaluate mathematical expressions involving multi-dimensional arrays efficiently. Since its introduction, it has been one of the most used CPU and GPU mathematical compilers - especially in the machine learning community - and has shown steady performance improvements. Theano is being actively and continuously developed since 2008, mu…
▽ More
Theano is a Python library that allows to define, optimize, and evaluate mathematical expressions involving multi-dimensional arrays efficiently. Since its introduction, it has been one of the most used CPU and GPU mathematical compilers - especially in the machine learning community - and has shown steady performance improvements. Theano is being actively and continuously developed since 2008, multiple frameworks have been built on top of it and it has been used to produce many state-of-the-art machine learning models.
The present article is structured as follows. Section I provides an overview of the Theano software and its community. Section II presents the principal features of Theano and how to use them, and compares them with other similar projects. Section III focuses on recently-introduced functionalities and improvements. Section IV compares the performance of Theano against Torch7 and TensorFlow on several machine learning models. Section V discusses current limitations of Theano and potential ways of improving it.
△ Less
Submitted 9 May, 2016;
originally announced May 2016.
-
Theano: new features and speed improvements
Authors:
Frédéric Bastien,
Pascal Lamblin,
Razvan Pascanu,
James Bergstra,
Ian Goodfellow,
Arnaud Bergeron,
Nicolas Bouchard,
David Warde-Farley,
Yoshua Bengio
Abstract:
Theano is a linear algebra compiler that optimizes a user's symbolically-specified mathematical computations to produce efficient low-level implementations. In this paper, we present new features and efficiency improvements to Theano, and benchmarks demonstrating Theano's performance relative to Torch7, a recently introduced machine learning library, and to RNNLM, a C++ library targeted at recurre…
▽ More
Theano is a linear algebra compiler that optimizes a user's symbolically-specified mathematical computations to produce efficient low-level implementations. In this paper, we present new features and efficiency improvements to Theano, and benchmarks demonstrating Theano's performance relative to Torch7, a recently introduced machine learning library, and to RNNLM, a C++ library targeted at recurrent neural networks.
△ Less
Submitted 23 November, 2012;
originally announced November 2012.
-
Deep Self-Taught Learning for Handwritten Character Recognition
Authors:
Frédéric Bastien,
Yoshua Bengio,
Arnaud Bergeron,
Nicolas Boulanger-Lewandowski,
Thomas Breuel,
Youssouf Chherawala,
Moustapha Cisse,
Myriam Côté,
Dumitru Erhan,
Jeremy Eustache,
Xavier Glorot,
Xavier Muller,
Sylvain Pannetier Lebeuf,
Razvan Pascanu,
Salah Rifai,
Francois Savard,
Guillaume Sicard
Abstract:
Recent theoretical and empirical work in statistical machine learning has demonstrated the importance of learning algorithms for deep architectures, i.e., function classes obtained by composing multiple non-linear transformations. Self-taught learning (exploiting unlabeled examples or examples from other distributions) has already been applied to deep learners, but mostly to show the advantage of…
▽ More
Recent theoretical and empirical work in statistical machine learning has demonstrated the importance of learning algorithms for deep architectures, i.e., function classes obtained by composing multiple non-linear transformations. Self-taught learning (exploiting unlabeled examples or examples from other distributions) has already been applied to deep learners, but mostly to show the advantage of unlabeled examples. Here we explore the advantage brought by {\em out-of-distribution examples}. For this purpose we developed a powerful generator of stochastic variations and noise processes for character images, including not only affine transformations but also slant, local elastic deformations, changes in thickness, background images, grey level changes, contrast, occlusion, and various types of noise. The out-of-distribution examples are obtained from these highly distorted images or by including examples of object classes different from those in the target test set. We show that {\em deep learners benefit more from out-of-distribution examples than a corresponding shallow learner}, at least in the area of handwritten character recognition. In fact, we show that they beat previously published results and reach human-level performance on both handwritten digit classification and 62-class handwritten character recognition.
△ Less
Submitted 18 September, 2010;
originally announced September 2010.
-
Parking functions, labeled trees and DCJ sorting scenarios
Authors:
Aida Ouangraoua,
Anne Bergeron
Abstract:
In genome rearrangement theory, one of the elusive questions raised in recent years is the enumeration of rearrangement scenarios between two genomes. This problem is related to the uniform generation of rearrangement scenarios, and the derivation of tests of statistical significance of the properties of these scenarios. Here we give an exact formula for the number of double-cut-and-join (DCJ) r…
▽ More
In genome rearrangement theory, one of the elusive questions raised in recent years is the enumeration of rearrangement scenarios between two genomes. This problem is related to the uniform generation of rearrangement scenarios, and the derivation of tests of statistical significance of the properties of these scenarios. Here we give an exact formula for the number of double-cut-and-join (DCJ) rearrangement scenarios of co-tailed genomes. We also construct effective bijections between the set of scenarios that sort a cycle and well studied combinatorial objects such as parking functions and labeled trees.
△ Less
Submitted 13 March, 2009;
originally announced March 2009.
