We describe a single convolutional neural network architecture that, given a sentence, outputs a host of language processing predictions: part-of-speech tags, chunks, named entity tags, semantic roles, semantically similar words and the likelihood that the sentence makes sense (grammatically and semantically) using a language model. The entire network is trained jointly on all these tasks using weight-sharing, an instance of multitask learning. All the tasks use labeled data except the language model which is learnt from unlabeled text and represents a novel form of semi-supervised learning for the shared tasks. We show how both multitask learning and semi-supervised learning improve the generalization of the shared tasks, resulting in state-of-the-art-performance.
%0 Conference Paper
%1 collobert2008unified
%A Collobert, Ronan
%A Weston, Jason
%B ICML
%C New York, NY, USA
%D 2008
%I ACM
%K language network neural nlp
%P 160--167
%R 10.1145/1390156.1390177
%T A Unified Architecture for Natural Language Processing: Deep Neural Networks with Multitask Learning
%U http://doi.acm.org/10.1145/1390156.1390177
%X We describe a single convolutional neural network architecture that, given a sentence, outputs a host of language processing predictions: part-of-speech tags, chunks, named entity tags, semantic roles, semantically similar words and the likelihood that the sentence makes sense (grammatically and semantically) using a language model. The entire network is trained jointly on all these tasks using weight-sharing, an instance of multitask learning. All the tasks use labeled data except the language model which is learnt from unlabeled text and represents a novel form of semi-supervised learning for the shared tasks. We show how both multitask learning and semi-supervised learning improve the generalization of the shared tasks, resulting in state-of-the-art-performance.
%@ 978-1-60558-205-4
@inproceedings{collobert2008unified,
abstract = {We describe a single convolutional neural network architecture that, given a sentence, outputs a host of language processing predictions: part-of-speech tags, chunks, named entity tags, semantic roles, semantically similar words and the likelihood that the sentence makes sense (grammatically and semantically) using a language model. The entire network is trained jointly on all these tasks using weight-sharing, an instance of multitask learning. All the tasks use labeled data except the language model which is learnt from unlabeled text and represents a novel form of semi-supervised learning for the shared tasks. We show how both multitask learning and semi-supervised learning improve the generalization of the shared tasks, resulting in state-of-the-art-performance.},
acmid = {1390177},
added-at = {2016-12-02T06:53:49.000+0100},
address = {New York, NY, USA},
author = {Collobert, Ronan and Weston, Jason},
biburl = {https://www.bibsonomy.org/bibtex/2ef98bb68b338e7a9ccc69f43493954b1/thoni},
booktitle = {ICML},
doi = {10.1145/1390156.1390177},
interhash = {3f9fd54a89bf8be67e9de4c766dcba83},
intrahash = {ef98bb68b338e7a9ccc69f43493954b1},
isbn = {978-1-60558-205-4},
keywords = {language network neural nlp},
location = {Helsinki, Finland},
numpages = {8},
pages = {160--167},
publisher = {ACM},
timestamp = {2017-05-16T09:50:09.000+0200},
title = {A Unified Architecture for Natural Language Processing: Deep Neural Networks with Multitask Learning},
url = {http://doi.acm.org/10.1145/1390156.1390177},
year = 2008
}