Controlling Output Length in Neural Encoder-Decoders
Authors:
- Yuta KikuchiYKYuta Kikuchi
- Graham NeubigGNGraham Neubig
- Ryohei SasanoRSRyohei Sasano
- Hiroya TakamuraHTHiroya Takamura
- Manabu OkumuraMOManabu Okumura
Preprint
Abstract
Neural encoder-decoder models have shown great success in many sequence generation tasks. However, previous work has not investigated situations in which we would like to control the length of encoder-decoder outputs. This capability is crucial for applications such as text summarization, in which we have to generate concise summaries with a desired length. In this paper, we propose methods for controlling the output sequence length for neural encoder-decoder models: two decoding-based methods and two learning-based methods. Results show that our learning-based methods have the capability to control length without degrading summary quality in a summarization task.Keywords
All Related Versions
This publication has 8 references indexed in Scilit:
- Incorporating Copying Mechanism in Sequence-to-Sequence Learning2016
- Abstractive Sentence Summarization with Attentive Recurrent Neural Networks2016
- Controlling Politeness in Neural Machine Translation via Side Constraints2016
- Effective Approaches to Attention-based Neural Machine Translation2015
- A Neural Attention Model for Abstractive Sentence Summarization2015
- Recurrent Continuous Translation Models2013
- Whole-sentence exponential language models: a vehicle for linguistic-statistical integrationComputer Speech & Language, 2001
- Headline generation based on statistical translation2000