KG Lab WS 2022 23 Task Scientific Event Series Completion and Announcement Crawler

Jump to bottom

Wolfgang Fahl edited this page Oct 20, 2022 · 33 revisions

Introduction

ConfIDent Project

CEUR-WS

Conference Corpus

Task

Starting Point: CEUR-WS Volume e.g. Volume 2065 (for all 3000+ Volumes)
in Wikidata KG navigate from Proceedings e.g. Proceedings of Workshops and Tutorials of the 9th International Conference on Knowledge Capture to Workshop Workshops and Tutorials of the 9th International Conference on Knowledge Capture
Find Conference + Series
- e.g. via DBLP K-CAP 2017 → DBLP K-CAP Series
- e.g. Conference Corpus RESTFul Query for K-CAP
- e.g. K-CAP Conference Wikidata(Scholia) Entry
Find (past) Announcements
- e.g. via K-CAP 2021 Wikidata entry → K-CAP 2021 Website
- previous K-CAP 2019 Wikidata entry → K-CAP 2019 Website
- previous K-CAP 2017 Wikidata entry → K-CAP 2017 Website
Guess Future Announcement - Find Trigger
- e.g. https://k-cap2023.org/ versus https://www.k-cap.org/2023
- e.g. Internet Search for K-CAP 2023 → Twitter message on K-CAP 2023 → Anna Lisa Gentile on Twitter → Anna Lisa Gentile on Scholia
Apply Event Series Completion as per past KG Lab results
- e.g. create Wikidata/ConfIDent/OpenResearch Entry for Event (K-CAP 2023 in our example)

Research Questions

How often can the algorithm be successfully applied?
Which obstacles have to be overcome to increase the success rate?
How can the success rate be improved by overcoming obstacles?

Technology Stack

Python
SPARQL backend, justpy frontend
existing libraries available e.g. for crawling (see e.g. Wiki-CFP crawler for ConferenceCorpus)
correctness check by statistical analysis - likelihood for correctness depending on length of series so far