Syntax Vector Learning using correspondence for Natural Language Understanding

📌 [Paper]

Authors: Hyein Seo, Sangkeun Jung, Taewook Hwang, Hyunji Kim, Yoon-Hyung Roh

Institution: Chungnam National University (Intelligent Software Laboratory)

How to cite

@ARTICLE{9448015,
    author={Seo, Hyein and Jung, Sangkeun and Hwang, Taewook and Kim, Hyunji and Roh, Yoon-Hyung},
    journal={IEEE Access},
    title={Syntax Vector Learning Using Correspondence for Natural Language Understanding},
    year={2021},
    volume={9},
    number={},
    pages={84067-84078},
    doi={10.1109/ACCESS.2021.3087271}}
}

📗 Dataset statistics

• Datasets
  • Weather, Navi, Rest : In-house
  • ATIS
  • SNIPS
  • Sim-M, Sim-R
  • NLUE

Section	Description
APPENDIX B	Graphs between cosine similarity and syntax similarity
APPENDIX C	Syntax search performance
APPENDIX D	Correlations between cosine similarity and syntax similarity

📊 APPENDIX B

This appendix provides additional graphs of the relationship between similarities in Weather, Navi, SNIPS, SIm-M, SIm-R, and NLUE datasets.

🔢 APPENDIX C

This appendix reports the syntax search scores for every task and model that we proposed in this paper. Table shows the predicted results according to the distances in the vector space.

📈 APPENDIX D

This appendix contains correlations between cosine similarity and syntax similarity results achieved on more datasets: Weather, Navi, SNIPS, Sim-M, Sim-R, NLUE.

Contact

Please reachout to [email protected] for any questions.