File Download
Supplementary
-
Citations:
- Appears in Collections:
Conference Paper: Learning distributed sentence vectors with bi-directional 3D convolutions
Title | Learning distributed sentence vectors with bi-directional 3D convolutions |
---|---|
Authors | |
Issue Date | 2020 |
Publisher | International Committee on Computational Linguistics. |
Citation | Proceedings of the 28th International Conference on Computational Linguistics (COLING 2020), Virtual Conference, Barcelona, Spain, 8-13 December 2020, p. 6820–6830 How to Cite? |
Abstract | We propose to learn distributed sentence representation using text’s visual features as input. Different from the existing methods that render the words or characters of a sentence into images separately, we further fold these images into a 3-dimensional sentence tensor. Then, multiple 3-dimensional convolutions with different lengths (the third dimension) are applied to the sentence tensor, which act as bi-gram, tri-gram, quad-gram, and even five-gram detectors jointly. Similar to the Bi-LSTM, these n-gram detectors learn both forward and backward distributional semantic knowledge from the sentence tensor. That is, the proposed model using bi-directional convolutions to learn text embedding according to the semantic order of words. The feature maps from the two directions are concatenated for final sentence embedding learning. Our model involves only a single-layer of convolution which makes it easy and fast to train. Finally, we evaluate the sentence embeddings on several downstream Natural Language Processing (NLP) tasks, which demonstrate a surprisingly excellent performance of the proposed model. |
Description | Session Poster25 - Machine Learning and Language Modelling. |
Persistent Identifier | http://hdl.handle.net/10722/294918 |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Liu, B | - |
dc.contributor.author | Wang, L | - |
dc.contributor.author | Yin, G | - |
dc.date.accessioned | 2020-12-21T11:50:24Z | - |
dc.date.available | 2020-12-21T11:50:24Z | - |
dc.date.issued | 2020 | - |
dc.identifier.citation | Proceedings of the 28th International Conference on Computational Linguistics (COLING 2020), Virtual Conference, Barcelona, Spain, 8-13 December 2020, p. 6820–6830 | - |
dc.identifier.uri | http://hdl.handle.net/10722/294918 | - |
dc.description | Session Poster25 - Machine Learning and Language Modelling. | - |
dc.description.abstract | We propose to learn distributed sentence representation using text’s visual features as input. Different from the existing methods that render the words or characters of a sentence into images separately, we further fold these images into a 3-dimensional sentence tensor. Then, multiple 3-dimensional convolutions with different lengths (the third dimension) are applied to the sentence tensor, which act as bi-gram, tri-gram, quad-gram, and even five-gram detectors jointly. Similar to the Bi-LSTM, these n-gram detectors learn both forward and backward distributional semantic knowledge from the sentence tensor. That is, the proposed model using bi-directional convolutions to learn text embedding according to the semantic order of words. The feature maps from the two directions are concatenated for final sentence embedding learning. Our model involves only a single-layer of convolution which makes it easy and fast to train. Finally, we evaluate the sentence embeddings on several downstream Natural Language Processing (NLP) tasks, which demonstrate a surprisingly excellent performance of the proposed model. | - |
dc.language | eng | - |
dc.publisher | International Committee on Computational Linguistics. | - |
dc.relation.ispartof | The 28th International Conference on Computational Linguistics (COLING 2020) | - |
dc.rights | This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. | - |
dc.title | Learning distributed sentence vectors with bi-directional 3D convolutions | - |
dc.type | Conference_Paper | - |
dc.identifier.email | Yin, G: gyin@hku.hk | - |
dc.identifier.authority | Yin, G=rp00831 | - |
dc.description.nature | published_or_final_version | - |
dc.identifier.hkuros | 320598 | - |
dc.identifier.spage | 6820 | - |
dc.identifier.epage | 6830 | - |