arxiv_id
stringlengths
10
10
title
stringlengths
8
162
summary
stringlengths
228
1.92k
pdf
stringlengths
14
14
authors
listlengths
1
451
categories
listlengths
1
6
primary_category
stringclasses
40 values
published
timestamp[s]date
2015-02-08 13:57:59
2024-01-08 18:47:34
updated
timestamp[s]date
2015-03-02 20:44:12
2026-06-12 08:15:52
references
null
content
stringlengths
8.83k
951k
1605.09090
Learning Natural Language Inference using Bidirectional LSTM model and Inner-Attention
In this paper, we proposed a sentence encoding-based model for recognizing text entailment. In our approach, the encoding of sentence is a two-stage process. Firstly, average pooling was used over word-level bidirectional LSTM (biLSTM) to generate a first-stage sentence representation. Secondly, attention mechanism was...
1605.09090.pdf
[ "Yang Liu", "Chengjie Sun", "Lei Lin", "Xiaolong Wang" ]
[ "cs.CL" ]
cs.CL
2016-05-30T02:47:35
2016-05-30T02:47:35
null
## Learning Natural Language Inference using Bidirectional LSTM model and Inner-Attention Yang Liu, Chengjie Sun, Lei Lin and Xiaolong Wang Harbin Institute of Technology, Harbin, P.R.China { yliu,cjsun,linl,wangxl } @insun.hit.edu.cn ## Abstract In this paper, we proposed a sentence encoding-based model for recog...
1605.08803
Density estimation using Real NVP
Unsupervised learning of probabilistic models is a central yet challenging problem in machine learning. Specifically, designing models with tractable learning, sampling, inference and evaluation is crucial in solving this task. We extend the space of such models using real-valued non-volume preserving (real NVP) transf...
1605.08803.pdf
[ "Laurent Dinh", "Jascha Sohl-Dickstein", "Samy Bengio" ]
[ "cs.LG", "cs.AI", "cs.NE", "stat.ML" ]
cs.LG
2016-05-27T21:24:32
2017-02-27T23:21:10
null
## DENSITY ESTIMATION USING REAL NVP Laurent Dinh ∗ Montreal Institute for Learning Algorithms University of Montreal Montreal, QC H3T1J4 Jascha Sohl-Dickstein Google Brain Samy Bengio Google Brain ## ABSTRACT Unsupervised learning of probabilistic models is a central yet challenging problem in machine learning. Sp...
1605.07725
Adversarial Training Methods for Semi-Supervised Text Classification
Adversarial training provides a means of regularizing supervised learning algorithms while virtual adversarial training is able to extend supervised learning algorithms to the semi-supervised setting. However, both methods require making small perturbations to numerous entries of the input vector, which is inappropriat...
1605.07725.pdf
[ "Takeru Miyato", "Andrew M. Dai", "Ian Goodfellow" ]
[ "stat.ML", "cs.LG" ]
stat.ML
2016-05-25T04:25:45
2021-11-16T07:16:21
null
## ADVERSARIAL TRAINING METHODS FOR SEMI-SUPERVISED TEXT CLASSIFICATION Takeru Miyato 1,2 ∗ , Andrew M Dai 2 , Ian Goodfellow 3 takeru.miyato@gmail.com, adai@google.com, ian@openai.com 1 Preferred Networks, Inc., ATR Cognitive Mechanisms Laboratories, Kyoto University 2 Google Brain 3 OpenAI ## ABSTRACT Adversar...
1605.07678
An Analysis of Deep Neural Network Models for Practical Applications
Since the emergence of Deep Neural Networks (DNNs) as a prominent technique in the field of computer vision, the ImageNet classification challenge has played a major role in advancing the state-of-the-art. While accuracy figures have steadily increased, the resource utilisation of winning models has not been properly t...
1605.07678.pdf
[ "Alfredo Canziani", "Adam Paszke", "Eugenio Culurciello" ]
[ "cs.CV" ]
cs.CV
2016-05-24T22:36:02
2017-04-14T23:40:21
null
## AN ANALYSIS OF DEEP NEURAL NETWORK MODELS FOR PRACTICAL APPLICATIONS ## Alfredo Canziani & Eugenio Culurciello Weldon School of Biomedical Engineering Purdue University { canziani,euge } @purdue.edu ## Adam Paszke Faculty of Mathematics, Informatics and Mechanics University of Warsaw a.paszke@students.mimu...
1605.07427
Hierarchical Memory Networks
Memory networks are neural networks with an explicit memory component that can be both read and written to by the network. The memory is often addressed in a soft way using a softmax function, making end-to-end training with backpropagation possible. However, this is not computationally scalable for applications which ...
1605.07427.pdf
[ "Sarath Chandar", "Sungjin Ahn", "Hugo Larochelle", "Pascal Vincent", "Gerald Tesauro", "Yoshua Bengio" ]
[ "stat.ML", "cs.CL", "cs.LG", "cs.NE" ]
stat.ML
2016-05-24T12:48:19
2016-05-24T12:48:19
null
## Hierarchical Memory Networks Sarath Chandar ∗ 1 , Sungjin Ahn 1 , Hugo Larochelle 2 , 4 , Pascal Vincent 1 , 4 , 3 1 , 4 Gerald Tesauro , Yoshua Bengio 1 Université de Montréal, Canada. 2 Twitter Cortex, USA. 3 IBM Watson Research Center, USA. 4 CIFAR, Canada. ## Abstract Memory networks are neural networks wit...
1605.07683
Learning End-to-End Goal-Oriented Dialog
Traditional dialog systems used in goal-oriented applications require a lot of domain-specific handcrafting, which hinders scaling up to new domains. End-to-end dialog systems, in which all components are trained from the dialogs themselves, escape this limitation. But the encouraging success recently obtained in chit-...
1605.07683.pdf
[ "Antoine Bordes", "Y-Lan Boureau", "Jason Weston" ]
[ "cs.CL" ]
cs.CL
2016-05-24T23:09:58
2017-03-30T23:02:22
null
## LEARNING END-TO-END GOAL-ORIENTED DIALOG Antoine Bordes, Y-Lan Boureau & Jason Weston Facebook AI Research New York, USA {abordes, ylan, jase}@fb.com ## ABSTRACT Traditional dialog systems used in goal-oriented applications require a lot of domain-specific handcrafting, which hinders scaling up to new domains...
1605.06431
Residual Networks Behave Like Ensembles of Relatively Shallow Networks
In this work we propose a novel interpretation of residual networks showing that they can be seen as a collection of many paths of differing length. Moreover, residual networks seem to enable very deep networks by leveraging only the short paths during training. To support this observation, we rewrite residual networks...
1605.06431.pdf
[ "Andreas Veit", "Michael Wilber", "Serge Belongie" ]
[ "cs.CV", "cs.AI", "cs.LG", "cs.NE" ]
cs.CV
2016-05-20T16:44:03
2016-10-27T00:43:58
null
## Residual Networks Behave Like Ensembles of Relatively Shallow Networks Andreas Veit Michael Wilber Serge Belongie Department of Computer Science & Cornell Tech Cornell University {av443, mjw285, sjb344}@cornell.edu ## Abstract In this work we propose a novel interpretation of residual networks showing that t...
1605.04711
Ternary Weight Networks
We present a memory and computation efficient ternary weight networks (TWNs) - with weights constrained to +1, 0 and -1. The Euclidian distance between full (float or double) precision weights and the ternary weights along with a scaling factor is minimized in training stage. Besides, a threshold-based ternary function...
1605.04711.pdf
[ "Fengfu Li", "Bin Liu", "Xiaoxing Wang", "Bo Zhang", "Junchi Yan" ]
[ "cs.CV" ]
cs.CV
2016-05-16T10:21:25
2022-11-20T14:21:55
null
## TERNARY WEIGHT NETWORKS Fengfu Li 1 † , Bin Liu 2 † , Xiaoxing Wang 2 , Bo Zhang 1 ∗ , Junchi Yan 2 ∗ 1 Institute of Applied Math., AMSS, CAS, Beijing, China lifengfu12@mails.ucas.ac.cn, b.zhang@amt.ac.cn 2 MOE Key Lab of Artificial Intelligence, Shanghai Jiao Tong University, Shanghai, China {binliu\_sjtu, fig...
1604.06778
Benchmarking Deep Reinforcement Learning for Continuous Control
Recently, researchers have made significant progress combining the advances in deep learning for learning feature representations with reinforcement learning. Some notable examples include training agents to play Atari games based on raw pixel data and to acquire advanced manipulation skills using raw sensory inputs. H...
1604.06778.pdf
[ "Yan Duan", "Xi Chen", "Rein Houthooft", "John Schulman", "Pieter Abbeel" ]
[ "cs.LG", "cs.AI", "cs.RO" ]
cs.LG
2016-04-22T18:57:24
2016-05-27T19:25:59
null
## Benchmarking Deep Reinforcement Learning for Continuous Control Yan Duan † Xi Chen † Rein Houthooft † ‡ John Schulman † § † Pieter Abbeel ROCKYDUAN@EECS.BERKELEY.EDU C.XI@EECS.BERKELEY.EDU REIN.HOUTHOOFT@UGENT.BE JOSCHU@EECS.BERKELEY.EDU PABBEEL@CS.BERKELEY.EDU † University of California, Berkeley, Department of...
1604.06174
Training Deep Nets with Sublinear Memory Cost
We propose a systematic approach to reduce the memory consumption of deep neural network training. Specifically, we design an algorithm that costs O(sqrt(n)) memory to train a n layer network, with only the computational cost of an extra forward pass per mini-batch. As many of the state-of-the-art models hit the upper ...
1604.06174.pdf
[ "Tianqi Chen", "Bing Xu", "Chiyuan Zhang", "Carlos Guestrin" ]
[ "cs.LG" ]
cs.LG
2016-04-21T04:15:27
2016-04-22T19:21:36
null
## Training Deep Nets with Sublinear Memory Cost Tianqi Chen 1 , Bing Xu 2 , Chiyuan Zhang 3 , and Carlos Guestrin 1 1 Unveristy of Washington 2 Dato. Inc 3 Massachusetts Institute of Technology ## Abstract We propose a systematic approach to reduce the memory consumption of deep neural network training. Specifical...
1604.03168
Hardware-oriented Approximation of Convolutional Neural Networks
High computational complexity hinders the widespread usage of Convolutional Neural Networks (CNNs), especially in mobile devices. Hardware accelerators are arguably the most promising approach for reducing both execution time and power consumption. One of the most important steps in accelerator development is hardware-...
1604.03168.pdf
[ "Philipp Gysel", "Mohammad Motamedi", "Soheil Ghiasi" ]
[ "cs.CV" ]
cs.CV
2016-04-11T22:43:21
2016-10-20T15:09:39
null
## HARDWARE-ORIENTED APPROXIMATION OF CONVOLUTIONAL NEURAL NETWORKS ## Philipp Gysel, Mohammad Motamedi & Soheil Ghiasi Department of Electrical and Computer Engineering University of California, Davis Davis, CA 95616, USA { pmgysel,mmotamedi,ghiasi } @ucdavis.edu ## ABSTRACT High computational complexity hin...
1604.00289
Building Machines That Learn and Think Like People
Recent progress in artificial intelligence (AI) has renewed interest in building systems that learn and think like people. Many advances have come from using deep neural networks trained end-to-end in tasks such as object recognition, video games, and board games, achieving performance that equals or even beats humans ...
1604.00289.pdf
[ "Brenden M. Lake", "Tomer D. Ullman", "Joshua B. Tenenbaum", "Samuel J. Gershman" ]
[ "cs.AI", "cs.CV", "cs.LG", "cs.NE", "stat.ML" ]
cs.AI
2016-04-01T15:37:57
2016-11-02T17:26:50
null
In press at Behavioral and Brain Sciences . ## Building Machines That Learn and Think Like People Brenden M. Lake, 1 Tomer D. Ullman, 2 , 4 Joshua B. Tenenbaum, 2 , 4 and Samuel J. Gershman 3 , 4 1 Center for Data Science, New York University 2 Department of Brain and Cognitive Sciences, MIT 3 Department of Psychol...
1603.09025
Recurrent Batch Normalization
We propose a reparameterization of LSTM that brings the benefits of batch normalization to recurrent neural networks. Whereas previous works only apply batch normalization to the input-to-hidden transformation of RNNs, we demonstrate that it is both possible and beneficial to batch-normalize the hidden-to-hidden transi...
1603.09025.pdf
[ "Tim Cooijmans", "Nicolas Ballas", "César Laurent", "Çağlar Gülçehre", "Aaron Courville" ]
[ "cs.LG" ]
cs.LG
2016-03-30T02:57:20
2017-02-28T00:59:42
null
## RECURRENT BATCH NORMALIZATION Tim Cooijmans, Nicolas Ballas, César Laurent, Ça˘ glar Gülçehre & Aaron Courville MILA - Université de Montréal firstname.lastname@umontreal.ca ## ABSTRACT We propose a reparameterization of LSTM that brings the benefits of batch normalization to recurrent neural networks. Wher...
1603.08983
Adaptive Computation Time for Recurrent Neural Networks
This paper introduces Adaptive Computation Time (ACT), an algorithm that allows recurrent neural networks to learn how many computational steps to take between receiving an input and emitting an output. ACT requires minimal changes to the network architecture, is deterministic and differentiable, and does not add any n...
1603.08983.pdf
[ "Alex Graves" ]
[ "cs.NE" ]
cs.NE
2016-03-29T22:09:00
2017-02-21T16:21:21
null
## Adaptive Computation Time for Recurrent Neural Networks Alex Graves Google DeepMind gravesa@google.com ## Abstract This paper introduces Adaptive Computation Time (ACT), an algorithm that allows recurrent neural networks to learn how many computational steps to take between receiving an input and emitting an outp...
1603.06147
A Character-Level Decoder without Explicit Segmentation for Neural Machine Translation
The existing machine translation systems, whether phrase-based or neural, have relied almost exclusively on word-level modelling with explicit segmentation. In this paper, we ask a fundamental question: can neural machine translation generate a character sequence without any explicit segmentation? To answer this questi...
1603.06147.pdf
[ "Junyoung Chung", "Kyunghyun Cho", "Yoshua Bengio" ]
[ "cs.CL", "cs.LG" ]
cs.CL
2016-03-19T21:35:04
2016-06-21T01:12:22
null
## A Character-Level Decoder without Explicit Segmentation for Neural Machine Translation Junyoung Chung Universit´ e de Montr´ eal junyoung.chung@umontreal.ca ## Abstract The existing machine translation systems, whether phrase-based or neural, have relied almost exclusively on word-level modelling with explicit ...
1603.05279
XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networks
We propose two efficient approximations to standard convolutional neural networks: Binary-Weight-Networks and XNOR-Networks. In Binary-Weight-Networks, the filters are approximated with binary values resulting in 32x memory saving. In XNOR-Networks, both the filters and the input to convolutional layers are binary. XNO...
1603.05279.pdf
[ "Mohammad Rastegari", "Vicente Ordonez", "Joseph Redmon", "Ali Farhadi" ]
[ "cs.CV" ]
cs.CV
2016-03-16T20:56:21
2016-08-02T20:59:14
null
## XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networks Mohammad Rastegari † , Vicente Ordonez † , Joseph Redmon ∗ , Ali Farhadi †∗ Allen Institute for AI † , University of Washington ∗ { mohammadr,vicenteor } @allenai.org { pjreddie,ali } @cs.washington.edu Abstract. We propose two efficien...
1603.05027
Identity Mappings in Deep Residual Networks
Deep residual networks have emerged as a family of extremely deep architectures showing compelling accuracy and nice convergence behaviors. In this paper, we analyze the propagation formulations behind the residual building blocks, which suggest that the forward and backward signals can be directly propagated from one ...
1603.05027.pdf
[ "Kaiming He", "Xiangyu Zhang", "Shaoqing Ren", "Jian Sun" ]
[ "cs.CV", "cs.LG" ]
cs.CV
2016-03-16T10:53:56
2016-07-25T15:18:32
null
## Identity Mappings in Deep Residual Networks Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun Microsoft Research Abstract Deep residual networks [1] have emerged as a family of extremely deep architectures showing compelling accuracy and nice convergence behaviors. In this paper, we analyze the propagation fo...
1603.04779
Revisiting Batch Normalization For Practical Domain Adaptation
Deep neural networks (DNN) have shown unprecedented success in various computer vision applications such as image classification and object detection. However, it is still a common annoyance during the training phase, that one has to prepare at least thousands of labeled images to fine-tune a network to a specific doma...
1603.04779.pdf
[ "Yanghao Li", "Naiyan Wang", "Jianping Shi", "Jiaying Liu", "Xiaodi Hou" ]
[ "cs.CV", "cs.LG" ]
cs.CV
2016-03-15T17:44:32
2016-11-08T06:11:30
null
## REVISITING BATCH NORMALIZATION FOR PRACTICAL DOMAIN ADAPTATION Yanghao Li † , Naiyan Wang ‡ , Jianping Shi ⋄ , Jiaying Liu † , Xiaodi Hou ‡ † Institute of Computer Science and Technology, Peking University ‡ TuSimple ⋄ SenseTime lyttonhao@pku.edu.cn winsty@gmail.com shijianping5000@gmail.com liujiaying@pku.edu....
1603.04467
TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems
TensorFlow is an interface for expressing machine learning algorithms, and an implementation for executing such algorithms. A computation expressed using TensorFlow can be executed with little or no change on a wide variety of heterogeneous systems, ranging from mobile devices such as phones and tablets up to large-sca...
1603.04467.pdf
[ "Martín Abadi", "Ashish Agarwal", "Paul Barham", "Eugene Brevdo", "Zhifeng Chen", "Craig Citro", "Greg S. Corrado", "Andy Davis", "Jeffrey Dean", "Matthieu Devin", "Sanjay Ghemawat", "Ian Goodfellow", "Andrew Harp", "Geoffrey Irving", "Michael Isard", "Yangqing Jia", "Rafal Jozefowic...
[ "cs.DC", "cs.LG" ]
cs.DC
2016-03-14T20:50:20
2016-03-16T16:57:12
null
## TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems (Preliminary White Paper, November 9, 2015) Mart´ ın Abadi, Ashish Agarwal, Paul Barham, Eugene Brevdo, Zhifeng Chen, Craig Citro, Greg S. Corrado, Andy Davis, Jeffrey Dean, Matthieu Devin, Sanjay Ghemawat, Ian Goodfellow, Andrew Harp, G...
1603.01360
Neural Architectures for Named Entity Recognition
State-of-the-art named entity recognition systems rely heavily on hand-crafted features and domain-specific knowledge in order to learn effectively from the small, supervised training corpora that are available. In this paper, we introduce two new neural architectures---one based on bidirectional LSTMs and conditional ...
1603.01360.pdf
[ "Guillaume Lample", "Miguel Ballesteros", "Sandeep Subramanian", "Kazuya Kawakami", "Chris Dyer" ]
[ "cs.CL" ]
cs.CL
2016-03-04T06:36:29
2016-04-07T15:09:36
null
## Neural Architectures for Named Entity Recognition Guillaume Lample ♠ Miguel Ballesteros ♣♠ Sandeep Subramanian ♠ Kazuya Kawakami ♠ Chris Dyer ♠ ♠ Carnegie Mellon University ♣ NLP Group, Pompeu Fabra University { glample,sandeeps,kkawakam,cdyer } @cs.cmu.edu, miguel.ballesteros@upf.edu ## Abstract State-of-the...
1603.01025
Convolutional Neural Networks using Logarithmic Data Representation
Recent advances in convolutional neural networks have considered model complexity and hardware efficiency to enable deployment onto embedded systems and mobile devices. For example, it is now well-known that the arithmetic operations of deep networks can be encoded down to 8-bit fixed-point without significant deterior...
1603.01025.pdf
[ "Daisuke Miyashita", "Edward H. Lee", "Boris Murmann" ]
[ "cs.NE", "cs.LG" ]
cs.NE
2016-03-03T08:51:52
2016-03-17T03:32:30
null
## Convolutional Neural Networks using Logarithmic Data Representation ## Daisuke Miyashita Stanford University, Stanford, CA 94305 USA Toshiba, Kawasaki, Japan ## Edward H. Lee Stanford University, Stanford, CA 94305 USA ## Boris Murmann Stanford University, Stanford, CA 94305 USA ## Abstract Recent advances i...
1602.07868
Weight Normalization: A Simple Reparameterization to Accelerate Training of Deep Neural Networks
We present weight normalization: a reparameterization of the weight vectors in a neural network that decouples the length of those weight vectors from their direction. By reparameterizing the weights in this way we improve the conditioning of the optimization problem and we speed up convergence of stochastic gradient d...
1602.07868.pdf
[ "Tim Salimans", "Diederik P. Kingma" ]
[ "cs.LG", "cs.AI", "cs.NE" ]
cs.LG
2016-02-25T10:13:45
2016-06-04T01:21:52
null
## Weight Normalization: A Simple Reparameterization to Accelerate Training of Deep Neural Networks Tim Salimans OpenAI tim@openai.com Diederik P. Kingma OpenAI dpkingma@openai.com ## Abstract We present weight normalization : a reparameterization of the weight vectors in a neural network that decouples the length...
1602.07360
SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
Recent research on deep neural networks has focused primarily on improving accuracy. For a given accuracy level, it is typically possible to identify multiple DNN architectures that achieve that accuracy level. With equivalent accuracy, smaller DNN architectures offer at least three advantages: (1) Smaller DNNs require...
1602.07360.pdf
[ "Forrest N. Iandola", "Song Han", "Matthew W. Moskewicz", "Khalid Ashraf", "William J. Dally", "Kurt Keutzer" ]
[ "cs.CV", "cs.AI" ]
cs.CV
2016-02-24T00:09:45
2016-11-04T21:26:08
null
## SQUEEZENET: ALEXNET-LEVEL ACCURACY WITH 50X FEWER PARAMETERS AND &lt; 0.5MB MODEL SIZE Forrest N. Iandola 1 , Song Han 2 , Matthew W. Moskewicz 1 , Khalid Ashraf 1 , William J. Dally 2 , Kurt Keutzer 1 1 DeepScale ∗ &amp;UCBerkeley 2 Stanford University { forresti, moskewcz, kashraf, keutzer } @eecs.berkeley.edu...
1602.07261
Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning
Very deep convolutional networks have been central to the largest advances in image recognition performance in recent years. One example is the Inception architecture that has been shown to achieve very good performance at relatively low computational cost. Recently, the introduction of residual connections in conjunct...
1602.07261.pdf
[ "Christian Szegedy", "Sergey Ioffe", "Vincent Vanhoucke", "Alex Alemi" ]
[ "cs.CV" ]
cs.CV
2016-02-23T18:44:39
2016-08-23T16:42:29
null
## Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning Christian Szegedy Google Inc. 1600 Amphitheatre Pkwy, Mountain View, CA szegedy@google.com Alex Alemi alemi@google.com ## Abstract Very deep convolutional networks have been central to the largest advances in image recognition pe...
1602.02867
Value Iteration Networks
We introduce the value iteration network (VIN): a fully differentiable neural network with a `planning module' embedded within. VINs can learn to plan, and are suitable for predicting outcomes that involve planning-based reasoning, such as policies for reinforcement learning. Key to our approach is a novel differentiab...
1602.02867.pdf
[ "Aviv Tamar", "Yi Wu", "Garrett Thomas", "Sergey Levine", "Pieter Abbeel" ]
[ "cs.AI", "cs.LG", "cs.NE", "stat.ML" ]
cs.AI
2016-02-09T05:44:36
2017-03-20T21:41:51
null
## Value Iteration Networks Aviv Tamar, Yi Wu, Garrett Thomas, Sergey Levine, and Pieter Abbeel Dept. of Electrical Engineering and Computer Sciences, UC Berkeley ## Abstract We introduce the value iteration network (VIN): a fully differentiable neural network with a 'planning module' embedded within. VINs can lear...
1602.02410
Exploring the Limits of Language Modeling
In this work we explore recent advances in Recurrent Neural Networks for large scale Language Modeling, a task central to language understanding. We extend current models to deal with two key challenges present in this task: corpora and vocabulary sizes, and complex, long term structure of language. We perform an exhau...
1602.02410.pdf
[ "Rafal Jozefowicz", "Oriol Vinyals", "Mike Schuster", "Noam Shazeer", "Yonghui Wu" ]
[ "cs.CL" ]
cs.CL
2016-02-07T19:11:17
2016-02-11T23:01:48
null
Rafal Jozefowicz Oriol Vinyals Mike Schuster Noam Shazeer Yonghui Wu Google Brain ## Abstract In this work we explore recent advances in Recurrent Neural Networks for large scale Language Modeling, a task central to language understanding. We extend current models to deal with two key challenges present in this task...
1602.01783
Asynchronous Methods for Deep Reinforcement Learning
We propose a conceptually simple and lightweight framework for deep reinforcement learning that uses asynchronous gradient descent for optimization of deep neural network controllers. We present asynchronous variants of four standard reinforcement learning algorithms and show that parallel actor-learners have a stabili...
1602.01783.pdf
[ "Volodymyr Mnih", "Adrià Puigdomènech Badia", "Mehdi Mirza", "Alex Graves", "Timothy P. Lillicrap", "Tim Harley", "David Silver", "Koray Kavukcuoglu" ]
[ "cs.LG" ]
cs.LG
2016-02-04T18:38:41
2016-06-16T16:38:45
null
## Asynchronous Methods for Deep Reinforcement Learning Volodymyr Mnih 1 Adrià Puigdomènech Badia 1 Mehdi Mirza 1 , 2 Alex Graves 1 Tim Harley 1 Timothy P. Lillicrap 1 David Silver 1 Koray Kavukcuoglu 1 1 Google DeepMind 2 Montreal Institute for Learning Algorithms (MILA), University of Montreal ## Abstract We pro...
1602.01137
A Dual Embedding Space Model for Document Ranking
A fundamental goal of search engines is to identify, given a query, documents that have relevant text. This is intrinsically difficult because the query and the document may use different vocabulary, or the document may contain query words without being relevant. We investigate neural word embeddings as a source of evi...
1602.01137.pdf
[ "Bhaskar Mitra", "Eric Nalisnick", "Nick Craswell", "Rich Caruana" ]
[ "cs.IR" ]
cs.IR
2016-02-02T22:23:18
2016-02-02T22:23:18
null
## A Dual Embedding Space Model for Document Ranking Bhaskar Mitra Microsoft Cambridge, UK bmitra@microsoft.com Eric Nalisnick University of California Irvine, USA enalisni@uci.edu ## ABSTRACT A fundamental goal of search engines is to identify, given a query, documents that have relevant text. This is intrinsically...
1602.00367
Efficient Character-level Document Classification by Combining Convolution and Recurrent Layers
Document classification tasks were primarily tackled at word level. Recent research that works with character-level inputs shows several benefits over word-level approaches such as natural incorporation of morphemes and better handling of rare words. We propose a neural network architecture that utilizes both convoluti...
1602.00367.pdf
[ "Yijun Xiao", "Kyunghyun Cho" ]
[ "cs.CL" ]
cs.CL
2016-02-01T02:53:41
2016-02-01T02:53:41
null
## Efficient Character-level Document Classification by Combining Convolution and Recurrent Layers Yijun Xiao Center for Data Sciences, New York University ryjxiao@nyu.edu ## Abstract Document classification tasks were primarily tackled at word level. Recent research that works with character-level inputs shows sev...
1601.06759
Pixel Recurrent Neural Networks
Modeling the distribution of natural images is a landmark problem in unsupervised learning. This task requires an image model that is at once expressive, tractable and scalable. We present a deep neural network that sequentially predicts the pixels in an image along the two spatial dimensions. Our method models the dis...
1601.06759.pdf
[ "Aaron van den Oord", "Nal Kalchbrenner", "Koray Kavukcuoglu" ]
[ "cs.CV", "cs.LG", "cs.NE" ]
cs.CV
2016-01-25T20:34:24
2016-08-19T14:10:16
null
A¨ aron van den Oord Nal Kalchbrenner Koray Kavukcuoglu Google DeepMind ## Abstract Modeling the distribution of natural images is a landmark problem in unsupervised learning. This task requires an image model that is at once expressive, tractable and scalable. We present a deep neural network that sequentially pred...
1601.01705
Learning to Compose Neural Networks for Question Answering
We describe a question answering model that applies to both images and structured knowledge bases. The model uses natural language strings to automatically assemble neural networks from a collection of composable modules. Parameters for these modules are learned jointly with network-assembly parameters via reinforcemen...
1601.01705.pdf
[ "Jacob Andreas", "Marcus Rohrbach", "Trevor Darrell", "Dan Klein" ]
[ "cs.CL", "cs.CV", "cs.NE" ]
cs.CL
2016-01-07T21:21:59
2016-06-07T23:25:51
null
## Learning to Compose Neural Networks for Question Answering ## Jacob Andreas and Marcus Rohrbach and Trevor Darrell and Dan Klein Department of Electrical Engineering and Computer Sciences University of California, Berkeley { jda,rohrbach,trevor,klein } @eecs.berkeley.edu ## Abstract We describe a question answe...
1601.00257
Modave Lectures on Applied AdS/CFT with Numerics
These lecture notes are intended to serve as an introduction to applied AdS/CFT with numerics for an audience of graduate students and others with little background in the subject. The presentation begins with a poor man's review of current status of quantum gravity, where AdS/CFT correspondence is believed to be the w...
1601.00257.pdf
[ "Minyong Guo", "Chao Niu", "Yu Tian", "Hongbao Zhang" ]
[ "gr-qc", "hep-th" ]
gr-qc
2016-01-03T08:10:29
2016-01-06T10:26:57
null
## Modave Lectures on Applied AdS/CFT with Numerics ∗ ## Minyong Guo Department of Physics, Beijing Normal University, Beijing, 100875, China minyongguo@mail.bnu.edu.cn ## Chao Niu School of Physics and Chemistry, Gwangju Institute of Science and Technology, Gwangju 500-712, Korea [chaoniu09@gmail.com](mailto:chao...
1512.05742
A Survey of Available Corpora for Building Data-Driven Dialogue Systems
During the past decade, several areas of speech and language understanding have witnessed substantial breakthroughs from the use of data-driven models. In the area of dialogue systems, the trend is less obvious, and most practical systems are still built through significant engineering and expert knowledge. Nevertheles...
1512.05742.pdf
[ "Iulian Vlad Serban", "Ryan Lowe", "Peter Henderson", "Laurent Charlin", "Joelle Pineau" ]
[ "cs.CL", "cs.AI", "cs.HC", "cs.LG", "stat.ML" ]
cs.CL
2015-12-17T19:52:39
2017-03-21T01:15:32
null
## A Survey of Available Corpora for Building Data-Driven Dialogue Systems ## Iulian Vlad Serban DIRO, Universit´ e de Montr´ eal 2920 chemin de la Tour, Montr´ eal, QC H3C 3J7, Canada ## Ryan Lowe Department of Computer Science, McGill University 3480 University st, Montr´ eal, QC H3A 0E9, Canada ## Peter Hender...
1512.04455
Memory-based control with recurrent neural networks
Partially observed control problems are a challenging aspect of reinforcement learning. We extend two related, model-free algorithms for continuous control -- deterministic policy gradient and stochastic value gradient -- to solve partially observed domains using recurrent neural networks trained with backpropagation t...
1512.04455.pdf
[ "Nicolas Heess", "Jonathan J Hunt", "Timothy P Lillicrap", "David Silver" ]
[ "cs.LG" ]
cs.LG
2015-12-14T18:44:48
2015-12-14T18:44:48
null
## Memory-based control with recurrent neural networks Nicolas Heess* Jonathan J Hunt* Timothy P Lillicrap David Silver Google Deepmind * These authors contributed equally. heess, jjhunt, countzero, davidsilver @ google.com ## Abstract Partially observed control problems are a challenging aspect of reinforcement ...
1512.03385
Deep Residual Learning for Image Recognition
Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced...
1512.03385.pdf
[ "Kaiming He", "Xiangyu Zhang", "Shaoqing Ren", "Jian Sun" ]
[ "cs.CV" ]
cs.CV
2015-12-10T19:51:55
2015-12-10T19:51:55
null
## Deep Residual Learning for Image Recognition Kaiming He Xiangyu Zhang Shaoqing Ren Microsoft Research { kahe, v-xiangz, v-shren, jiansun } @microsoft.com ## Abstract Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantiall...
1512.02167
Simple Baseline for Visual Question Answering
We describe a very simple bag-of-words baseline for visual question answering. This baseline concatenates the word features from the question and CNN features from the image to predict the answer. When evaluated on the challenging VQA dataset [2], it shows comparable performance to many recent approaches using recurren...
1512.02167.pdf
[ "Bolei Zhou", "Yuandong Tian", "Sainbayar Sukhbaatar", "Arthur Szlam", "Rob Fergus" ]
[ "cs.CV", "cs.CL" ]
cs.CV
2015-12-07T19:00:54
2015-12-15T05:17:49
null
## Simple Baseline for Visual Question Answering Bolei Zhou 1 , Yuandong Tian 2 , Sainbayar Sukhbaatar 2 , Arthur Szlam 2 , and Rob Fergus 2 1 Massachusetts Institute of Technology 2 Facebook AI Research ## Abstract We describe a very simple bag-of-words baseline for visual question answering. This baseline concate...
1512.00567
Rethinking the Inception Architecture for Computer Vision
Convolutional networks are at the core of most state-of-the-art computer vision solutions for a wide variety of tasks. Since 2014 very deep convolutional networks started to become mainstream, yielding substantial gains in various benchmarks. Although increased model size and computational cost tend to translate to imm...
1512.00567.pdf
[ "Christian Szegedy", "Vincent Vanhoucke", "Sergey Ioffe", "Jonathon Shlens", "Zbigniew Wojna" ]
[ "cs.CV" ]
cs.CV
2015-12-02T03:44:38
2015-12-11T20:27:50
null
## Rethinking the Inception Architecture for Computer Vision Christian Szegedy Google Inc. Vincent Vanhoucke Sergey Ioffe vanhoucke@google.com szegedy@google.com Zbigniew Wojna University College London zbigniewwojna@gmail.com larly high performance in the 2014 ILSVRC [16] classification challenge. One interesting ...
1511.08630
A C-LSTM Neural Network for Text Classification
Neural network models have been demonstrated to be capable of achieving remarkable performance in sentence and document modeling. Convolutional neural network (CNN) and recurrent neural network (RNN) are two mainstream architectures for such modeling tasks, which adopt totally different ways of understanding natural la...
1511.08630.pdf
[ "Chunting Zhou", "Chonglin Sun", "Zhiyuan Liu", "Francis C. M. Lau" ]
[ "cs.CL" ]
cs.CL
2015-11-27T11:44:17
2015-11-30T07:20:46
null
## A C-LSTM Neural Network for Text Classification Chunting Zhou 1 , Chonglin Sun 2 , Zhiyuan Liu 3 , Francis C.M. Lau 1 Department of Computer Science, The University of Hong Kong 1 School of Innovation Experiment, Dalian University of Technology 2 Department of Computer Science and Technology, Tsinghua University, ...
1511.06939
Session-based Recommendations with Recurrent Neural Networks
We apply recurrent neural networks (RNN) on a new domain, namely recommender systems. Real-life recommender systems often face the problem of having to base recommendations only on short session-based data (e.g. a small sportsware website) instead of long user histories (as in the case of Netflix). In this situation th...
1511.06939.pdf
[ "Balázs Hidasi", "Alexandros Karatzoglou", "Linas Baltrunas", "Domonkos Tikk" ]
[ "cs.LG", "cs.IR", "cs.NE" ]
cs.LG
2015-11-21T23:42:59
2016-03-29T14:52:58
null
## SESSION-BASED RECOMMENDATIONS WITH RECURRENT NEURAL NETWORKS Bal´ azs Hidasi ∗ Gravity R&amp;D Inc. Budapest, Hungary balazs.hidasi@gravityrd.com Linas Baltrunas † Netflix Los Gatos, CA, USA lbaltrunas@netflix.com Alexandros Karatzoglou Telefonica Research Barcelona, Spain alexk@tid.es Domonkos Tikk Gravity R&amp...
1511.06807
Adding Gradient Noise Improves Learning for Very Deep Networks
Deep feedforward and recurrent networks have achieved impressive results in many perception and language processing applications. This success is partially attributed to architectural innovations such as convolutional and long short-term memory networks. The main motivation for these architectural innovations is that t...
1511.06807.pdf
[ "Arvind Neelakantan", "Luke Vilnis", "Quoc V. Le", "Ilya Sutskever", "Lukasz Kaiser", "Karol Kurach", "James Martens" ]
[ "stat.ML", "cs.LG" ]
stat.ML
2015-11-21T01:11:29
2015-11-21T01:11:29
null
## ADDING GRADIENT NOISE IMPROVES LEARNING FOR VERY DEEP NETWORKS ## Arvind Neelakantan ∗ , Luke Vilnis ∗ College of Information and Computer Sciences University of Massachusetts Amherst { arvind,luke } @cs.umass.edu Quoc V. Le, Ilya Sutskever, Lukasz Kaiser, Karol Kurach Google Brain { qvl,ilyasu,lukaszkaiser,k...
1511.06789
The Unreasonable Effectiveness of Noisy Data for Fine-Grained Recognition
Current approaches for fine-grained recognition do the following: First, recruit experts to annotate a dataset of images, optionally also collecting more structured data in the form of part annotations and bounding boxes. Second, train a model utilizing this data. Toward the goal of solving fine-grained recognition, we...
1511.06789.pdf
[ "Jonathan Krause", "Benjamin Sapp", "Andrew Howard", "Howard Zhou", "Alexander Toshev", "Tom Duerig", "James Philbin", "Li Fei-Fei" ]
[ "cs.CV" ]
cs.CV
2015-11-20T22:40:30
2016-10-18T18:35:31
null
## The Unreasonable Effectiveness of Noisy Data for Fine-Grained Recognition Jonathan Krause 1 ⋆ Benjamin Sapp 2 ⋆⋆ Andrew Howard 3 Howard Zhou 3 Alexander Toshev 3 Tom Duerig 3 James Philbin 2 ⋆⋆ Li Fei-Fei 1 1 Stanford University 2 Zoox 3 Google { jkrause,feifeili } @cs.stanford.edu { bensapp,james } @zoox.com { how...
1511.06488
Resiliency of Deep Neural Networks under Quantization
The complexity of deep neural network algorithms for hardware implementation can be much lowered by optimizing the word-length of weights and signals. Direct quantization of floating-point weights, however, does not show good performance when the number of bits assigned is small. Retraining of quantized networks has be...
1511.06488.pdf
[ "Wonyong Sung", "Sungho Shin", "Kyuyeon Hwang" ]
[ "cs.LG", "cs.NE" ]
cs.LG
2015-11-20T04:55:46
2016-01-07T13:50:22
null
## RESILIENCY OF DEEP NEURAL NETWORKS UNDER QUANTIZATION ## Wonyong Sung, Sungho Shin &amp; Kyuyeon Hwang Department of Electrical and Computer Engineering Seoul National University Seoul, 08826 Korea wysung@snu.ac.kr shshin@dsp.snu.ac.kr kyuyeon.hwang@gmail.com ## ABSTRACT The complexity of deep neural network al...
1511.06297
Conditional Computation in Neural Networks for faster models
Deep learning has become the state-of-art tool in many applications, but the evaluation and training of deep models can be time-consuming and computationally expensive. The conditional computation approach has been proposed to tackle this problem (Bengio et al., 2013; Davis & Arel, 2013). It operates by selectively act...
1511.06297.pdf
[ "Emmanuel Bengio", "Pierre-Luc Bacon", "Joelle Pineau", "Doina Precup" ]
[ "cs.LG" ]
cs.LG
2015-11-19T18:40:22
2016-01-07T22:41:10
null
## CONDITIONAL COMPUTATION IN NEURAL NETWORKS FOR FASTER MODELS Emmanuel Bengio, Pierre-Luc Bacon, Joelle Pineau &amp; Doina Precup School of Computer Science McGill University Montreal, Canada { ebengi,pbacon,jpineau,dprecup } @cs.mcgill.ca ## ABSTRACT Deep learning has become the state-of-art tool in many appli...
1511.06279
Neural Programmer-Interpreters
We propose the neural programmer-interpreter (NPI): a recurrent and compositional neural network that learns to represent and execute programs. NPI has three learnable components: a task-agnostic recurrent core, a persistent key-value program memory, and domain-specific encoders that enable a single NPI to operate in m...
1511.06279.pdf
[ "Scott Reed", "Nando de Freitas" ]
[ "cs.LG", "cs.NE" ]
cs.LG
2015-11-19T17:49:32
2016-02-29T11:12:36
null
## NEURAL PROGRAMMER-INTERPRETERS Scott Reed &amp; Nando de Freitas Google DeepMind London, UK scott.ellison.reed@gmail.com nandodefreitas@google.com ## ABSTRACT We propose the neural programmer-interpreter (NPI): a recurrent and compositional neural network that learns to represent and execute programs. NPI has t...
1511.06434
Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
In recent years, supervised learning with convolutional networks (CNNs) has seen huge adoption in computer vision applications. Comparatively, unsupervised learning with CNNs has received less attention. In this work we hope to help bridge the gap between the success of CNNs for supervised learning and unsupervised lea...
1511.06434.pdf
[ "Alec Radford", "Luke Metz", "Soumith Chintala" ]
[ "cs.LG", "cs.CV" ]
cs.LG
2015-11-19T22:50:32
2016-01-07T23:09:39
null
## UNSUPERVISED REPRESENTATION LEARNING WITH DEEP CONVOLUTIONAL GENERATIVE ADVERSARIAL NETWORKS Alec Radford &amp; Luke Metz indico Research Boston, MA { alec,luke } @indico.io Soumith Chintala Facebook AI Research New York, NY soumith@fb.com ## ABSTRACT In recent years, supervised learning with convolutional netw...
1511.06342
Actor-Mimic: Deep Multitask and Transfer Reinforcement Learning
The ability to act in multiple environments and transfer previous knowledge to new situations can be considered a critical aspect of any intelligent agent. Towards this goal, we define a novel method of multitask and transfer learning that enables an autonomous agent to learn how to behave in multiple tasks simultaneou...
1511.06342.pdf
[ "Emilio Parisotto", "Jimmy Lei Ba", "Ruslan Salakhutdinov" ]
[ "cs.LG" ]
cs.LG
2015-11-19T20:17:27
2016-02-22T19:59:40
null
## ACTOR-MIMIC DEEP MULTITASK AND TRANSFER REINFORCEMENT LEARNING ## Emilio Parisotto, Jimmy Ba, Ruslan Salakhutdinov Department of Computer Science University of Toronto Toronto, Ontario, Canada { eparisotto,jimmy,rsalakhu } @cs.toronto.edu ## ABSTRACT The ability to act in multiple environments and transfer prev...
1511.05756
Image Question Answering using Convolutional Neural Network with Dynamic Parameter Prediction
We tackle image question answering (ImageQA) problem by learning a convolutional neural network (CNN) with a dynamic parameter layer whose weights are determined adaptively based on questions. For the adaptive parameter prediction, we employ a separate parameter prediction network, which consists of gated recurrent uni...
1511.05756.pdf
[ "Hyeonwoo Noh", "Paul Hongsuck Seo", "Bohyung Han" ]
[ "cs.CV", "cs.CL", "cs.LG" ]
cs.CV
2015-11-18T12:30:57
2015-11-18T12:30:57
null
## Image Question Answering using Convolutional Neural Network with Dynamic Parameter Prediction Hyeonwoo Noh Bohyung Han Department of Computer Science and Engineering, POSTECH, Korea Paul Hongsuck Seo { hyeonwoonoh , hsseo, bhhan } @postech.ac.kr ## Abstract We tackle image question answering (ImageQA) problem b...
1511.05234
Ask, Attend and Answer: Exploring Question-Guided Spatial Attention for Visual Question Answering
We address the problem of Visual Question Answering (VQA), which requires joint image and language understanding to answer a question about a given photograph. Recent approaches have applied deep image captioning methods based on convolutional-recurrent networks to this problem, but have failed to model spatial inferen...
1511.05234.pdf
[ "Huijuan Xu", "Kate Saenko" ]
[ "cs.CV", "cs.AI", "cs.CL", "cs.NE" ]
cs.CV
2015-11-17T01:00:04
2016-03-19T03:06:58
null
## Ask, Attend and Answer: Exploring Question-Guided Spatial Attention for Visual Question Answering Huijuan Xu and Kate Saenko Department of Computer Science, UMass Lowell, USA hxu1@cs.uml.edu , saenko@cs.uml.edu Abstract. Weaddress the problem of Visual Question Answering (VQA), which requires joint image and lang...
1511.04636
Deep Reinforcement Learning with a Natural Language Action Space
This paper introduces a novel architecture for reinforcement learning with deep neural networks designed to handle state and action spaces characterized by natural language, as found in text-based games. Termed a deep reinforcement relevance network (DRRN), the architecture represents action and state spaces with separ...
1511.04636.pdf
[ "Ji He", "Jianshu Chen", "Xiaodong He", "Jianfeng Gao", "Lihong Li", "Li Deng", "Mari Ostendorf" ]
[ "cs.AI", "cs.CL", "cs.LG" ]
cs.AI
2015-11-14T23:30:39
2016-06-08T05:58:34
null
## Deep Reinforcement Learning with a Natural Language Action Space † ∗ Department of Electrical Engineering, University of Washington, Seattle, WA 98195, USA { jvking, ostendor } @uw.edu Ji He ∗ , Jianshu Chen † , Xiaodong He † , Jianfeng Gao † , Lihong Li Li Deng † and Mari Ostendorf ∗ † Microsoft Research, Redmo...
1511.02274
Stacked Attention Networks for Image Question Answering
This paper presents stacked attention networks (SANs) that learn to answer natural language questions from images. SANs use semantic representation of a question as query to search for the regions in an image that are related to the answer. We argue that image question answering (QA) often requires multiple steps of re...
1511.02274.pdf
[ "Zichao Yang", "Xiaodong He", "Jianfeng Gao", "Li Deng", "Alex Smola" ]
[ "cs.LG", "cs.CL", "cs.CV", "cs.NE" ]
cs.LG
2015-11-07T00:43:32
2016-01-26T20:37:49
null
## Stacked Attention Networks for Image Question Answering Zichao Yang 1 , Xiaodong He 2 , Jianfeng Gao 2 , Li Deng 2 , Alex Smola 1 1 Carnegie Mellon University, 2 Microsoft Research, Redmond, WA 98052, USA zichaoy@cs.cmu.edu, { xiaohe, jfgao, deng } @microsoft.com, alex@smola.org ## Abstract This paper presents s...
1510.03009
Neural Networks with Few Multiplications
For most deep learning algorithms training is notoriously time consuming. Since most of the computation in training neural networks is typically spent on floating point multiplications, we investigate an approach to training that eliminates the need for most of these. Our method consists of two parts: First we stochast...
1510.03009.pdf
[ "Zhouhan Lin", "Matthieu Courbariaux", "Roland Memisevic", "Yoshua Bengio" ]
[ "cs.LG", "cs.NE" ]
cs.LG
2015-10-11T04:32:39
2016-02-26T05:24:30
null
## NEURAL NETWORKS WITH FEW MULTIPLICATIONS Zhouhan Lin Universit´ e de Montr´ eal Canada zhouhan.lin@umontreal.ca Roland Memisevic Universit´ e de Montr´ eal Canada roland.umontreal@gmail.com Matthieu Courbariaux Universit´ e de Montr´ eal Canada matthieu.courbariaux@gmail.com Yoshua Bengio Universit´ e de Montr´ ea...
1510.02675
Controlled Experiments for Word Embeddings
An experimental approach to studying the properties of word embeddings is proposed. Controlled experiments, achieved through modifications of the training corpus, permit the demonstration of direct relations between word properties and word vector direction and length. The approach is demonstrated using the word2vec CB...
1510.02675.pdf
[ "Benjamin J. Wilson", "Adriaan M. J. Schakel" ]
[ "cs.CL" ]
cs.CL
2015-10-09T14:03:33
2015-12-14T05:10:29
null
## Controlled Experiments for Word Embeddings Benjamin Wilson Lateral GmbH benjamin@lateral.io Adriaan M. J. Schakel NNLP adriaan.schakel@gmail.com December 15, 2015 ## Abstract Anexperimentalapproach to studying the properties of word embeddings is proposed. Controlled experiments, achieved through modifications ...
1510.01378
Batch Normalized Recurrent Neural Networks
Recurrent Neural Networks (RNNs) are powerful models for sequential data that have the potential to learn long-term dependencies. However, they are computationally expensive to train and difficult to parallelize. Recent work has shown that normalizing intermediate representations of neural networks can significantly im...
1510.01378.pdf
[ "César Laurent", "Gabriel Pereyra", "Philémon Brakel", "Ying Zhang", "Yoshua Bengio" ]
[ "stat.ML", "cs.LG", "cs.NE" ]
stat.ML
2015-10-05T21:45:31
2015-10-05T21:45:31
null
## Batch Normalized Recurrent Neural Networks C´ esar Laurent ∗ Universit´ e de Montr´ eal Phil´ emon Brakel Universit´ e de Montr´ eal Gabriel Pereyra ∗ University of Southern California Ying Zhang Universit´ e de Montr´ eal Yoshua Bengio † Universit´ e de Montr´ eal ## Abstract Recurrent Neural Networks (RNNs...
1510.00149
Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding
Neural networks are both computationally intensive and memory intensive, making them difficult to deploy on embedded systems with limited hardware resources. To address this limitation, we introduce "deep compression", a three stage pipeline: pruning, trained quantization and Huffman coding, that work together to reduc...
1510.00149.pdf
[ "Song Han", "Huizi Mao", "William J. Dally" ]
[ "cs.CV", "cs.NE" ]
cs.CV
2015-10-01T09:03:44
2016-02-15T06:25:40
null
## DEEP COMPRESSION: COMPRESSING DEEP NEURAL NETWORKS WITH PRUNING, TRAINED QUANTIZATION AND HUFFMAN CODING ## Song Han Stanford University, Stanford, CA 94305, USA songhan@stanford.edu ## Huizi Mao Tsinghua University, Beijing, 100084, China mhz12@mails.tsinghua.edu.cn ## William J. Dally Stanford University, St...
1509.03005
Compatible Value Gradients for Reinforcement Learning of Continuous Deep Policies
This paper proposes GProp, a deep reinforcement learning algorithm for continuous policies with compatible function approximation. The algorithm is based on two innovations. Firstly, we present a temporal-difference based method for learning the gradient of the value-function. Secondly, we present the deviator-actor-cr...
1509.03005.pdf
[ "David Balduzzi", "Muhammad Ghifary" ]
[ "cs.LG", "cs.AI", "cs.NE", "stat.ML" ]
cs.LG
2015-09-10T04:14:54
2015-09-10T04:14:54
null
## Compatible Value Gradients for Reinforcement Learning of Continuous Deep Policies ## David Balduzzi School of Mathematics and Statistics Victoria University of Wellington Wellington, New Zealand ## Muhammad Ghifary School of Engineering and Computer Science Victoria University of Wellington Wellington, New Zeala...
1509.02971
Continuous control with deep reinforcement learning
We adapt the ideas underlying the success of Deep Q-Learning to the continuous action domain. We present an actor-critic, model-free algorithm based on the deterministic policy gradient that can operate over continuous action spaces. Using the same learning algorithm, network architecture and hyper-parameters, our algo...
1509.02971.pdf
[ "Timothy P. Lillicrap", "Jonathan J. Hunt", "Alexander Pritzel", "Nicolas Heess", "Tom Erez", "Yuval Tassa", "David Silver", "Daan Wierstra" ]
[ "cs.LG", "stat.ML" ]
cs.LG
2015-09-09T23:01:36
2019-07-05T10:47:27
null
## CONTINUOUS CONTROL WITH DEEP REINFORCEMENT LEARNING Timothy P. Lillicrap , Jonathan J. Hunt , Alexander Pritzel, Nicolas Heess, ∗ ∗ Tom Erez, Yuval Tassa, David Silver &amp; Daan Wierstra Google Deepmind London, UK { countzero, jjhunt, apritzel, heess, etom, tassa, davidsilver, wierstra } @ google.com ## ABSTRA...
1508.06615
Character-Aware Neural Language Models
We describe a simple neural language model that relies only on character-level inputs. Predictions are still made at the word-level. Our model employs a convolutional neural network (CNN) and a highway network over characters, whose output is given to a long short-term memory (LSTM) recurrent neural network language mo...
1508.06615.pdf
[ "Yoon Kim", "Yacine Jernite", "David Sontag", "Alexander M. Rush" ]
[ "cs.CL", "cs.NE", "stat.ML" ]
cs.CL
2015-08-26T19:25:34
2015-12-01T22:59:24
null
Yoon Kim † ## Character-Aware Neural Language Models Yacine Jernite ∗ † School of Engineering and Applied Sciences Harvard University { yoonkim,srush } @seas.harvard.edu ## Abstract We describe a simple neural language model that relies only on character-level inputs. Predictions are still made at the word-level....
1508.05326
A large annotated corpus for learning natural language inference
Understanding entailment and contradiction is fundamental to understanding natural language, and inference about entailment and contradiction is a valuable testing ground for the development of semantic representations. However, machine learning research in this area has been dramatically limited by the lack of large-s...
1508.05326.pdf
[ "Samuel R. Bowman", "Gabor Angeli", "Christopher Potts", "Christopher D. Manning" ]
[ "cs.CL" ]
cs.CL
2015-08-21T16:17:01
2015-08-21T16:17:01
null
## A large annotated corpus for learning natural language inference Samuel R. Bowman ∗† sbowman@stanford.edu Christopher Potts ∗ cgpotts@stanford.edu Gabor Angeli †‡ angeli@stanford.edu Christopher D. Manning ∗†‡ manning@stanford.edu ∗ Stanford Linguistics † Stanford NLP Group ‡ Stanford Computer Science ## Abst...
1506.08909
The Ubuntu Dialogue Corpus: A Large Dataset for Research in Unstructured Multi-Turn Dialogue Systems
This paper introduces the Ubuntu Dialogue Corpus, a dataset containing almost 1 million multi-turn dialogues, with a total of over 7 million utterances and 100 million words. This provides a unique resource for research into building dialogue managers based on neural language models that can make use of large amounts o...
1506.08909.pdf
[ "Ryan Lowe", "Nissan Pow", "Iulian Serban", "Joelle Pineau" ]
[ "cs.CL", "cs.AI", "cs.LG", "cs.NE" ]
cs.CL
2015-06-30T00:37:09
2016-02-04T01:21:35
null
## The Ubuntu Dialogue Corpus: A Large Dataset for Research in Unstructured Multi-Turn Dialogue Systems Ryan Lowe ∗ * , Nissan Pow * , Iulian V. Serban † and Joelle Pineau * * † School of Computer Science, McGill University, Montreal, Canada Department of Computer Science and Operations Research, Universié de Montr...
1506.06724
Aligning Books and Movies: Towards Story-like Visual Explanations by Watching Movies and Reading Books
Books are a rich source of both fine-grained information, how a character, an object or a scene looks like, as well as high-level semantics, what someone is thinking, feeling and how these states evolve through a story. This paper aims to align books to their movie releases in order to provide rich descriptive explanat...
1506.06724.pdf
[ "Yukun Zhu", "Ryan Kiros", "Richard Zemel", "Ruslan Salakhutdinov", "Raquel Urtasun", "Antonio Torralba", "Sanja Fidler" ]
[ "cs.CV", "cs.CL" ]
cs.CV
2015-06-22T19:26:56
2015-06-22T19:26:56
null
## Aligning Books and Movies: Towards Story-like Visual Explanations by Watching Movies and Reading Books Yukun Zhu ∗ , 1 Ryan Kiros *,1 Richard Zemel 1 Ruslan Salakhutdinov 1 Raquel Urtasun 1 Antonio Torralba 2 Sanja Fidler 1 1 University of Toronto 2 Massachusetts Institute of Technology { yukun,rkiros,zemel,rsalak...
1506.02438
High-Dimensional Continuous Control Using Generalized Advantage Estimation
Policy gradient methods are an appealing approach in reinforcement learning because they directly optimize the cumulative reward and can straightforwardly be used with nonlinear function approximators such as neural networks. The two main challenges are the large number of samples typically required, and the difficulty...
1506.02438.pdf
[ "John Schulman", "Philipp Moritz", "Sergey Levine", "Michael Jordan", "Pieter Abbeel" ]
[ "cs.LG", "cs.RO", "eess.SY" ]
cs.LG
2015-06-08T11:12:48
2018-10-20T18:55:07
null
## HIGH-DIMENSIONAL CONTINUOUS CONTROL USING GENERALIZED ADVANTAGE ESTIMATION John Schulman, Philipp Moritz, Sergey Levine, Michael I. Jordan and Pieter Abbeel Department of Electrical Engineering and Computer Science University of California, Berkeley { joschu,pcmoritz,levine,jordan,pabbeel } @eecs.berkeley.edu ## ...
1506.02626
Learning both Weights and Connections for Efficient Neural Networks
Neural networks are both computationally intensive and memory intensive, making them difficult to deploy on embedded systems. Also, conventional networks fix the architecture before training starts; as a result, training cannot improve the architecture. To address these limitations, we describe a method to reduce the s...
1506.02626.pdf
[ "Song Han", "Jeff Pool", "John Tran", "William J. Dally" ]
[ "cs.NE", "cs.CV", "cs.LG" ]
cs.NE
2015-06-08T19:28:43
2015-10-30T23:29:27
null
## Learning both Weights and Connections for Efficient Neural Networks Song Han Stanford University songhan@stanford.edu John Tran NVIDIA johntran@nvidia.com Jeff Pool NVIDIA jpool@nvidia.com William J. Dally Stanford University NVIDIA dally@stanford.edu ## Abstract Neural networks are both computationally intensi...
1506.01186
Cyclical Learning Rates for Training Neural Networks
It is known that the learning rate is the most important hyper-parameter to tune for training deep neural networks. This paper describes a new method for setting the learning rate, named cyclical learning rates, which practically eliminates the need to experimentally find the best values and schedule for the global lea...
1506.01186.pdf
[ "Leslie N. Smith" ]
[ "cs.CV", "cs.LG", "cs.NE" ]
cs.CV
2015-06-03T09:54:31
2017-04-04T11:34:46
null
## Cyclical Learning Rates for Training Neural Networks Leslie N. Smith U.S. Naval Research Laboratory, Code 5514 4555 Overlook Ave., SW., Washington, D.C. 20375 leslie.smith@nrl.navy.mil Figure 1. Classification accuracy while training CIFAR-10. The red curve shows the result of training with one of the new learni...
1506.01066
Visualizing and Understanding Neural Models in NLP
While neural networks have been successfully applied to many NLP tasks the resulting vector-based models are very difficult to interpret. For example it's not clear how they achieve {\em compositionality}, building sentence meaning from the meanings of words and phrases. In this paper we describe four strategies for vi...
1506.01066.pdf
[ "Jiwei Li", "Xinlei Chen", "Eduard Hovy", "Dan Jurafsky" ]
[ "cs.CL" ]
cs.CL
2015-06-02T21:17:31
2016-01-08T18:10:22
null
## Visualizing and Understanding Neural Models in NLP Jiwei Li 1 , Xinlei Chen 2 , Eduard Hovy 2 and Dan Jurafsky 1 1 Computer Science Department, Stanford University, Stanford, CA 94305, USA 2 Language Technology Institute, Carnegie Mellon University, Pittsburgh, PA 15213, USA { jiweil,jurafsky } @stanford.edu ##...
1506.02488
On the Fuzzy Stability of an Affine Functional Equation
In this paper, we obtain the general solution of the following functional equation f(3x + y + z) + f(x + 3y + z) + f(x + y + 3z) + f(x) + f(y) + f(z) = 6f(x + y + z): We establish the Hyers-Ulam-Rassias stability of the above functional equation in the fuzzy normed spaces. Further we show the above functional equation ...
1506.02488.pdf
[ "Md. Nasiruzzaman" ]
[ "math.CA" ]
math.CA
2015-05-24T14:13:54
2015-05-24T14:13:54
null
## On the Fuzzy Stability of an Affine Functional Equation ## Md. Nasiruzzaman Department of Mathematics, Aligarh Muslim University, Aligarh 202002, India Email: nasir3489@gmail.com Abstract: In this paper, we obtain the general solution of the following functional equation <!-- formula-not-decoded --> We establi...
1505.00853
Empirical Evaluation of Rectified Activations in Convolutional Network
In this paper we investigate the performance of different types of rectified activation functions in convolutional neural network: standard rectified linear unit (ReLU), leaky rectified linear unit (Leaky ReLU), parametric rectified linear unit (PReLU) and a new randomized leaky rectified linear units (RReLU). We evalu...
1505.00853.pdf
[ "Bing Xu", "Naiyan Wang", "Tianqi Chen", "Mu Li" ]
[ "cs.LG", "cs.CV", "stat.ML" ]
cs.LG
2015-05-05T01:16:39
2015-11-27T06:58:14
null
## Empirical Evaluation of Rectified Activations in Convolution Network ## Bing Xu University of Alberta Naiyan Wang Hong Kong University of Science and Technology Tianqi Chen University of Washington Mu Li Carnegie Mellon University ## Abstract In this paper we investigate the performance of different types ...
1505.00521
Reinforcement Learning Neural Turing Machines - Revised
The Neural Turing Machine (NTM) is more expressive than all previously considered models because of its external memory. It can be viewed as a broader effort to use abstract external Interfaces and to learn a parametric model that interacts with them. The capabilities of a model can be extended by providing it with p...
1505.00521.pdf
[ "Wojciech Zaremba", "Ilya Sutskever" ]
[ "cs.LG" ]
cs.LG
2015-05-04T04:14:54
2016-01-12T06:35:48
null
## REINFORCEMENT LEARNING NEURAL TURING MACHINES - REVISED Wojciech Zaremba 1 , 2 New York University Facebook AI Research woj.zaremba@gmail.com Ilya Sutskever 2 Google Brain ilyasu@google.com ## ABSTRACT The Neural Turing Machine (NTM) is more expressive than all previously considered models because of its external...
1505.00387
Highway Networks
There is plenty of theoretical and empirical evidence that depth of neural networks is a crucial ingredient for their success. However, network training becomes more difficult with increasing depth and training of very deep networks remains an open problem. In this extended abstract, we introduce a new architecture des...
1505.00387.pdf
[ "Rupesh Kumar Srivastava", "Klaus Greff", "Jürgen Schmidhuber" ]
[ "cs.LG", "cs.NE" ]
cs.LG
2015-05-03T01:56:57
2015-11-03T18:15:15
null
Rupesh Kumar Srivastava Klaus Greff J¨ urgen Schmidhuber ## Highway Networks The Swiss AI Lab IDSIA Istituto Dalle Molle di Studi sull'Intelligenza Artificiale Universit` a della Svizzera italiana (USI) Scuola universitaria professionale della Svizzera italiana (SUPSI) Galleria 2, 6928 Manno-Lugano, Switzerland #...
1504.00702
End-to-End Training of Deep Visuomotor Policies
Policy search methods can allow robots to learn control policies for a wide range of tasks, but practical applications of policy search often require hand-engineered components for perception, state estimation, and low-level control. In this paper, we aim to answer the following question: does training the perception a...
1504.00702.pdf
[ "Sergey Levine", "Chelsea Finn", "Trevor Darrell", "Pieter Abbeel" ]
[ "cs.LG", "cs.CV", "cs.RO" ]
cs.LG
2015-04-02T22:23:51
2016-04-19T01:33:13
null
## End-to-End Training of Deep Visuomotor Policies Sergey Levine † Chelsea Finn † Trevor Darrell Pieter Abbeel Division of Computer Science University of California Berkeley, CA 94720-1776, USA † These authors contributed equally. Editor: Jan Peters svlevine@eecs.berkeley.edu cbfinn@eecs.berkeley.edu trevor@ee...
1504.00325
Microsoft COCO Captions: Data Collection and Evaluation Server
In this paper we describe the Microsoft COCO Caption dataset and evaluation server. When completed, the dataset will contain over one and a half million captions describing over 330,000 images. For the training and validation images, five independent human generated captions will be provided. To ensure consistency in e...
1504.00325.pdf
[ "Xinlei Chen", "Hao Fang", "Tsung-Yi Lin", "Ramakrishna Vedantam", "Saurabh Gupta", "Piotr Dollar", "C. Lawrence Zitnick" ]
[ "cs.CV", "cs.CL" ]
cs.CV
2015-04-01T18:13:43
2015-04-03T20:21:16
null
## Microsoft COCO Captions: Data Collection and Evaluation Server Xinlei Chen, Hao Fang, Tsung-Yi Lin, Ramakrishna Vedantam Saurabh Gupta, Piotr Doll´ ar, C. Lawrence Zitnick Abstract -In this paper we describe the Microsoft COCO Caption dataset and evaluation server. When completed, the dataset will contain over one...
1502.05698
Towards AI-Complete Question Answering: A Set of Prerequisite Toy Tasks
One long-term goal of machine learning research is to produce methods that are applicable to reasoning and natural language, in particular building an intelligent dialogue agent. To measure progress towards that goal, we argue for the usefulness of a set of proxy tasks that evaluate reading comprehension via question a...
1502.05698.pdf
[ "Jason Weston", "Antoine Bordes", "Sumit Chopra", "Alexander M. Rush", "Bart van Merriënboer", "Armand Joulin", "Tomas Mikolov" ]
[ "cs.AI", "cs.CL", "stat.ML" ]
cs.AI
2015-02-19T20:46:10
2015-12-31T13:08:14
null
## TOWARDS AI-COMPLETE QUESTION ANSWERING: A SET OF PREREQUISITE TOY TASKS Jason Weston, Antoine Bordes, Sumit Chopra, Alexander M. Rush, Bart van Merri¨ enboer, Armand Joulin &amp; Tomas Mikolov Facebook AI Research 770 Broadway New York, USA { jase,abordes,spchopra,tmikolov,sashar,bartvm } @fb.com ## ABSTRACT On...
1502.03167
Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Training Deep Neural Networks is complicated by the fact that the distribution of each layer's inputs changes during training, as the parameters of the previous layers change. This slows down the training by requiring lower learning rates and careful parameter initialization, and makes it notoriously hard to train mode...
1502.03167.pdf
[ "Sergey Ioffe", "Christian Szegedy" ]
[ "cs.LG" ]
cs.LG
2015-02-11T01:44:18
2015-03-02T20:44:12
null
## Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift Sergey Ioffe Google Inc., sioffe@google.com Christian Szegedy Google Inc., szegedy@google.com Using mini-batches of examples, as opposed to one example at a time, is helpful in several ways. First, the gradient of the loss...
1502.02251
From Pixels to Torques: Policy Learning with Deep Dynamical Models
Data-efficient learning in continuous state-action spaces using very high-dimensional observations remains a key challenge in developing fully autonomous systems. In this paper, we consider one instance of this challenge, the pixels to torques problem, where an agent must learn a closed-loop control policy from pixel i...
1502.02251.pdf
[ "Niklas Wahlström", "Thomas B. Schön", "Marc Peter Deisenroth" ]
[ "stat.ML", "cs.LG", "cs.RO", "eess.SY" ]
stat.ML
2015-02-08T13:57:59
2015-06-18T16:59:43
null
## From Pixels to Torques: Policy Learning with Deep Dynamical Models ## Niklas Wahlstr¨ om Division of Automatic Control, Link¨ oping University, Link¨ oping, Sweden ## Thomas B. Sch¨ on Department of Information Technology, Uppsala University, Sweden ## Marc Peter Deisenroth Department of Computing, Imperial Co...