Project Name: wikidata-autodesc
Wikitech Usernames of requestors: Kingsaint1989
Purpose: This project is intended to automatically generate missing descriptions for Wikidata items using existing statements of the items
Brief description: In our paper titled "Be Concise and Precise: Synthesizing Open-domain Entity Descriptions from Facts" published in the Proceedings of the Web Conference 2019 ( ArXiv link: https://arxiv.org/pdf/1904.07391.pdf ), we have developed a tool that can potentially be used for automatic generation of missing textual descriptions for millions of items in Wikidata. Our deep neural network model is developed in PyTorch, an open source deep learning framework. We would like to host this tool in the Cloud-VPS. We also plan to have a user interface where users can validate whether the generated descriptions are acceptable or not. For that, we need a web service and a web framework in place. Any Python-based web framework will be good.
How soon you are hoping this can be fulfilled: Within a quarter, but we may need to keep the tool online if it turns out to be effective for the long run.