Recurrent Nested Model for Sequence Generation
Lists
Only you see this
Abstract
Depth has been shown beneficial to neural network models. In this paper, we make an attempt to make the encoder-decoder model deeper for sequence generation. We propose a module that can be plugged into the middle between the encoder and decoder to increase the depth of the whole model. The proposed module follows a nested structure, which is divided into blocks with each block containing several recurrent transition steps. To reduce the training difficulty and preserve the necessary information for the decoder during transitions, inter-block connections and intra-block connections are constructed in our model. The inter-block connections provide the thought vectors from the current block to all the subsequent blocks. The intra-block connections connect all the hidden states entering the current block to the current transition step. The advantages of our model are illustrated on the image captioning and code captioning tasks.
BibTeX
@inproceedings{jiang2020recurrent,
author = {Wenhao Jiang and Lin Ma and Wei Lu},
title = {{Recurrent Nested Model for Sequence Generation}},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {2020},
doi = {10.1609/aaai.v34i07.6768},
}| Nested LSTMs | arXiv · 2018 |
| LSTM-in-LSTM for generating long descriptions of images | Comput. Vis. Media · 2016 |
| Recursion in Recursion: Two-Level Nested Recursion for Length Generalization with Scalability | NeurIPS · 2023 |
| CNN+CNN: Convolutional Decoders for Image Captioning | arXiv · 2018 |
| Deep Recurrent Models with Fast-Forward Connections for Neural Machine Translation | arXiv · 2016 |
| Recurrent Fusion Network for Image Captioning | arXiv · 2018 |