2018

Identifying and Controlling Important Neurons in Neural Machine Translation

Bau, Anthony, Belinkov, Yonatan, Sajjad, Hassan et al.

Understand

Neural machine translation (NMT) models learn representations containing substantial linguistic information.

However, it is not clear if such information is fully distributed or if some of it can be attributed to individual neurons.
We develop unsupervised methods for discovering important neurons in NMT models.
Our methods rely on the intuition that different models learn similar properties, and do not require any costly external supervision.

Built on

Distributed representations, simple recurrent networks, and grammatical structure
Jeffrey L. Elman · 1991
Earlier work this paper cites.
Compositional connectionism in cognitive science ii: the localist/distributed dimension
Ross W. Gayler and Simon D. Levy · 2011
Earlier work this paper cites.
A Simple, Fast, and Effective Reparameterization of IBM Model 2
Chris Dyer, Victor Chahuneau, and Noah A. Smith · 2013
Earlier work this paper cites.
Neural machine translation by jointly learning to align and translate
Original
Dzmitry Bahdanau, Kyunghyun Cho, and Yoshua Bengio · 2014
Earlier work this paper cites.
Visualizing and understanding convolutional networks
Matthew D Zeiler and Rob Fergus · 2014
Earlier work this paper cites.
Visualizing and understanding recurrent networks
Original
Andrej Karpathy, Justin Johnson, and Li Fei-Fei · 2015
Earlier work this paper cites.
Character-aware Neural Language Models
Original
Yoon Kim, Yacine Jernite, David Sontag, and Alexander M Rush · 2015
Earlier work this paper cites.
What’s in an Embedding? Analyzing Word Embeddings through Multilingual Evaluation
Arne Köhn · 2015
Earlier work this paper cites.
Fine-grained Analysis of Sentence Embeddings Using Auxiliary Prediction Tasks
Original
Yossi Adi, Einat Kermany, Yonatan Belinkov, Ofer Lavi, and Yoav Goldberg · 2016
Earlier work this paper cites.
Representation of linguistic form and function in recurrent neural networks
Original
Ákos Kádár, Grzegorz Chrupała, and Afra Alishahi · 2016
Earlier work this paper cites.

Similar

The Mythos of Model Interpretability
Zachary C Lipton · 2016
Cited alongside, same era.
Analyzing Linguistic Knowledge in Sequential Model of Sentence
Peng Qian, Xipeng Qiu, and Xuanjing Huang · 2016
Cited alongside, same era.
Neural machine translation of rare words with subword units
Rico Sennrich, Barry Haddow, and Alexandra Birch · 2016
Cited alongside, same era.
Why Neural Translations are the Right Length
Xing Shi, Kevin Knight, and Deniz Yuret · 2016
Cited alongside, same era.
Does String-Based Neural MT Learn Source Syntax?
Xing Shi, Inkit Padhi, and Kevin Knight · 2016
Cited alongside, same era.
Variational Neural Machine Translation
Biao Zhang, Deyi Xiong, jinsong su, Hong Duan, and Min Zhang · 2016
Cited alongside, same era.
Learning Deep Features for Discriminative Localization
B. Zhou, A. Khosla, Lapedriza. A., A. Oliva, and A. Torralba · 2016
Cited alongside, same era.
The United Nations Parallel Corpus v1.0
Michał Ziemski, Marcin Junczys-Dowmunt, and Bruno Pouliquen · 2016
Cited alongside, same era.
Network dissection: Quantifying interpretability of deep visual representations
David Bau, Bolei Zhou, Aditya Khosla, Aude Oliva, and Antonio Torralba · 2017
Cited alongside, same era.
What do Neural Machine Translation Models Learn about Morphology?
Yonatan Belinkov, Nadir Durrani, Fahim Dalvi, Hassan Sajjad, and James Glass · 2017
Cited alongside, same era.
Understanding Neural Networks through Representation Erasure
Original
Jiwei Li, Will Monroe, and Dan Jurafsky
Cited in the paper.
Convergent Learning: Do different neural networks learn the same representations?
Yixuan Li, Jason Yosinski, Jeff Clune, Hod Lipson, and John Hopcroft
Cited in the paper.

Then

Interpretation of Semantic Tweet Representations
J. Ganesh, Manish Gupta, and Vasudeva Varma · 2017
Later among the works it cites.
Convolutional Sequence to Sequence Learning
Jonas Gehring, Michael Auli, David Grangier, Denis Yarats, and Yann N. Dauphin · 2017
Later among the works it cites.
The representational geometry of word meanings acquired by neural machine translation models
Felix Hill, Kyunghyun Cho, Sébastien Jean, and Yoshua Bengio · 2017
Later among the works it cites.
Learning to generate reviews and discovering sentiment
Original
Alec Radford, Rafal Jozefowicz, and Ilya Sutskever · 2017
Later among the works it cites.
SVCCA: Singular Vector Canonical Correlation Analysis for Deep Learning Dynamics and Interpretability
Maithra Raghu, Justin Gilmer, Jason Yosinski, and Jascha Sohl-Dickstein · 2017
Later among the works it cites.
Attention is All you Need
Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Ł ukasz Kaiser, and Illia Polosukhin · 2017
Later among the works it cites.
Natural Language Multitasking: Analyzing and Improving Syntactic Saliency of Hidden Representations
Original
Gino Brunner, Yuyi Wang, Roger Wattenhofer, and Michael Weigelt · 2018
Closest in time.
On the importance of single directions for generalization
Ari S. Morcos, David G.T. Barrett, Neil C. Rabinowitz, and Matthew Botvinick · 2018
Closest in time.
Variational Recurrent Neural Machine Translation
Jinsong Su, Shan Wu, Deyi Xiong, Yaojie Lu, Xianpei Han, and Biao Zhang · 2018
Closest in time.
Revisiting the Importance of Individual Units in CNNs via Ablation
Original
Bolei Zhou, Yiyou Sun, David Bau, and Antonio Torralba · 2018
Closest in time.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…

Reading the bibliography…