ML p(r)ior | Efficient Character-level Document Classification by Combining Convolution and Recurrent Layers

Efficient Character-level Document Classification by Combining Convolution and Recurrent Layers

2016-02-01
1602.00367 | cs.CL
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 convolution and recurrent layers to efficiently encode character inputs. We validate the proposed model on eight large scale document classification tasks and compare with character-level convolution-only models. It achieves comparable performances with much less parameters.
PDF

Highlights - Most important sentences from the article

Login to like/save this paper, take notes and configure your recommendations

Related Articles

2017-06-12

The dominant sequence transduction models are based on complex recurrent or convolutional neural net… show more
PDF

Highlights - Most important sentences from the article

2015-11-04
1511.01432 | cs.LG

We present two approaches that use unlabeled data to improve sequence learning with recurrent networ… show more
PDF

Highlights - Most important sentences from the article

2016-08-25

Recent work has shown that convolutional networks can be substantially deeper, more accurate, and ef… show more
PDF

Highlights - Most important sentences from the article

2018-02-15

We introduce a new type of deep contextualized word representation that models both (1) complex char… show more
PDF

Highlights - Most important sentences from the article

2017-05-08

The prevalent approach to sequence to sequence learning maps an input sequence to a variable length … show more
PDF

Highlights - Most important sentences from the article

2016-06-06

The dominant approach for many NLP tasks are recurrent neural networks, in particular LSTMs, and con… show more
PDF

Highlights - Most important sentences from the article

2016-07-21

Training state-of-the-art, deep neural networks is computationally expensive. One way to reduce the … show more
PDF

Highlights - Most important sentences from the article

2018-03-04

For most deep learning practitioners, sequence modeling is synonymous with recurrent networks. Yet r… show more
PDF

Highlights - Most important sentences from the article

2016-02-07

In this work we explore recent advances in Recurrent Neural Networks for large scale Language Modeli… show more
PDF

Highlights - Most important sentences from the article

2015-09-04
1509.01626 | cs.LG

This article offers an empirical exploration on the use of character-level convolutional networks (C… show more
PDF

Highlights - Most important sentences from the article

2015-08-26

We describe a simple neural language model that relies only on character-level inputs. Predictions a… show more
PDF

Highlights - Most important sentences from the article

2016-03-04
1603.01354 | cs.LG

State-of-the-art sequence labeling systems traditionally require large amounts of task-specific know… show more
PDF

Highlights - Most important sentences from the article

2015-12-08

We show that an end-to-end deep learning approach can be used to recognize either English or Mandari… show more
PDF

Highlights - Most important sentences from the article

2016-12-23

The pre-dominant approach to language modeling to date is based on recurrent neural networks. Their … show more
PDF

Highlights - Most important sentences from the article

2014-12-24

We present an attention-based model for recognizing multiple objects in images. The proposed model i… show more
PDF

Highlights - Most important sentences from the article