IPTC EXTRA Rules Based Classification for News
Download as PPTX, PDF0 likes491 views
Google's €150 million Digital News Initiative Fund aims to foster innovation in European news organizations, with initial funding of €27 million distributed to projects across 23 countries. The document details the development of an open-source, rules-based multilingual news classification system, designed to improve tagging efficiency and accuracy. A budget of €50,000 is allocated for project development, and milestones include recruiting a team and evaluating existing open-source frameworks.
IPTC EXTRA Rules Based Classification for News
- 1. EXTRA Stuart Myles * Associated Press * 7th March 2016 © 2016 IPTC (www.iptc.org) All rights reserved https://flic.kr/p/81HXTG
- 2. Google DNI • Google’s €150 million Digital News Initiative fund – Stimulate innovation among European news organizations – https://www.digitalnewsinitiative.com/fund/ • Multiple rounds – first in October 2015, next in June 16 – First funding of €27 million to projects in 23 countries – http://googlepolicyeurope.blogspot.gr/2016/02/digital-news- initiative-first-funding_24.html © 2016 IPTC (www.iptc.org) All rights reserved
- 3. EXTRA EXTraction Rules Apparatus • Open source, rules-based, multilingual news classification – Rules in two languages for IPTC Media Topics • Rules based > Hand tagging – Consistent, performant and scalable • Rules based > Statistical tagging – Statistical approaches require numerous annotated examples – EXTRA will allow publishers to rapidly adapt to breaking news and low-frequency topics – Avoid problems with ambiguity (“Police Can’t Stop Gambling”) – Precisely distinguish between similar topics, which are more challenging for statistical approaches. © 2016 IPTC (www.iptc.org) All rights reserved 3
- 4. EXTRA Deliverables • Rules based engine • Formal specification of the classification rules language • Two sets of extraction rules – Tagging with IPTC Media Topics – Drive adoption and consistency © 2010 IPTC (www.iptc.org) All rights reserved 4
- 5. EXTRA Budget • Developer €35,000 • Linguist €10,000 • Project Manager €5,000 • Total - €50,000 © 2010 IPTC (www.iptc.org) All rights reserved 5
- 6. EXTRA Milestones • Recruit EXTRA Team • Evaluate existing open source projects and frameworks • Create first EXTRA release • Develop EXTRA software and rule sets • First non-core open source contributor • First production deployment © 2010 IPTC (www.iptc.org) All rights reserved 6
- 7. Apply for next round? • Apply in May 2016 for June round • Candidate applications – EXTRA – but no specific feedback – Sport Identifiers © 2016 IPTC (www.iptc.org) All rights reserved 7