Order Matters: Sequence to sequence for sets

Sequences have become first class citizens in supervised learning thanks to the resurgence of recurrent neural networks. Many complex tasks that require mapping from or to a sequence of observations can now be formulated with the sequence-to-sequence (seq2seq) framework which employs the chain rule to efficiently represent the joint probability of sequences. In many cases, however, variable sized inputs and/or outputs might not be naturally expressed as sequences. For instance, it is not clear how to input a set of numbers into a model where the task is to sort them; similarly, we do not know how to organize outputs when they correspond to random variables and the task is to model their unknown joint probability. In this paper, we first show using various examples that the order in which we organize input and/or output data matters significantly when learning an underlying model. We then discuss an extension of the seq2seq framework that goes beyond sequences and handles input sets in a principled way. In addition, we propose a loss which, by searching over possible orders during training, deals with the lack of structure of output sets. We show empirical evidence of our claims regarding ordering, and on the modifications to the seq2seq framework on benchmark language modeling and parsing tasks, as well as two artificial tasks -- sorting numbers and estimating the joint probability of unknown graphical models.

[1]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[2]  Gökhan BakIr,et al.  Predicting Structured Data , 2008 .

[3]  Andrew Zisserman,et al.  Advances in Neural Information Processing Systems (NIPS) , 2007 .

[4]  Alex Graves,et al.  Supervised Sequence Labelling with Recurrent Neural Networks , 2012, Studies in Computational Intelligence.

[5]  Christopher D. Manning,et al.  Learning Continuous Phrase Representations and Syntactic Parsing with Recursive Neural Networks , 2010 .

[6]  Andrew L. Maas,et al.  Word-level Acoustic Modeling with Convolutional Vector Regression , 2012 .

[7]  Tara N. Sainath,et al.  Deep Neural Networks for Acoustic Modeling in Speech Recognition , 2012 .

[8]  Phil Blunsom,et al.  Recurrent Continuous Translation Models , 2013, EMNLP.

[9]  Yoshua Bengio,et al.  Learning Phrase Representations using RNN Encoder–Decoder for Statistical Machine Translation , 2014, EMNLP.

[10]  Wojciech Zaremba,et al.  Learning to Execute , 2014, ArXiv.

[11]  Alex Graves,et al.  Neural Turing Machines , 2014, ArXiv.

[12]  Quoc V. Le,et al.  Sequence to Sequence Learning with Neural Networks , 2014, NIPS.

[13]  Wojciech Zaremba,et al.  Recurrent Neural Network Regularization , 2014, ArXiv.

[14]  Geoffrey E. Hinton,et al.  Grammar as a Foreign Language , 2014, NIPS.

[15]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[16]  Wei Xu,et al.  Deep Captioning with Multimodal Recurrent Neural Networks (m-RNN) , 2014, ICLR.

[17]  Jason Weston,et al.  Memory Networks , 2014, ICLR.

[18]  Navdeep Jaitly,et al.  Pointer Networks , 2015, NIPS.

[19]  Samy Bengio,et al.  Show and tell: A neural image caption generator , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[20]  Trevor Darrell,et al.  Long-term recurrent convolutional networks for visual recognition and description , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[21]  Wojciech Zaremba,et al.  Reinforcement Learning Neural Turing Machines , 2015, ArXiv.

[22]  Yoshua Bengio,et al.  Neural Machine Translation by Jointly Learning to Align and Translate , 2014, ICLR.

[23]  武田 一哉,et al.  Recurrent Neural Networkに基づく日常生活行動認識 , 2016 .

[24]  Yoshua Bengio,et al.  End-to-end attention-based large vocabulary speech recognition , 2015, 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[25]  Xinyun Chen.  Under Review as a Conference Paper at Iclr 2017 Delving into Transferable Adversarial Ex- Amples and Black-box Attacks , 2016 .

[26]  Tapani Raiko,et al.  International Conference on Learning Representations (ICLR) , 2016 .

[27]  Omer Levy,et al.  Published as a conference paper at ICLR 2018 S IMULATING A CTION D YNAMICS WITH N EURAL P ROCESS N ETWORKS , 2018 .