The AKBC 2019 Workshop on Federated KBs and Open Knowledge Network will focus on the challenges in creating a semantic information infrastructure for powering the next generation of AI-enabled, "smart" applications in diverse domains.
The AKBC Workshop on Federated KBs and the Open Knowledge Network will focus on the challenges in creating a semantic information infrastructure for powering the next generation of AI-enabled, "smart" applications in diverse domains. The Open Knowledge Network (OKN) builds upon technologies in Big Data, Semantic Web, Natural Language Processing, and Machine Learning to discover and create linkages among entities across disconnected knowledge-bases. Similar to the Internet, which began as an attempt to link files but became a major information infrastructure, the OKN begins by creating linkages among entities, but has the potential to become the enabling infrastructure for next generation AI-enabled, "smart" applications.
The US National Science Foundation recently announced the creation of the Open Knowledge Network as a Convergence Accelerator project, to help advance the progression from data to information to knowledge. Building the OKN requires addressing "horizontal" challenges that apply to all domains, such as the underlying representation of facts and relationships; query and inference services that perform reasoning tasks with the data; tracking provenance; authorized access capabilities; and the like. It also requires addressing "vertical" challenges that may be specific to data from particular topical domains such as, say, socioeconomics, geosciences, education, smart health, finance, manufacturing, etc. A recent workshop report articulated the research and development challenges in developing the Open Knowledge Network, as well as the potential benefits it offers for driving innovation in scientific discovery and economic growth.
Samuel J. Klein (MIT and Harvard): Enhancing current OKNs: Federating Wikidata & the Underlay Project
François Scharffe (Columbia University and The Data Chefs): Open Knowledge Embeddings (abstract pdf)
Vicki Tardif (Google): Challenges in Knowledge Graph Construction for Question Answering
Daniel Garijo (USC ISI): Towards an Open Knowledge Network of Reusable Software Descriptions
1:30-1:35: Brief welcome
1:35-2:00: Invited Talk #1 (20min talk + 5m questions):
2:00-2:25: Invited Talk #2 (20min talk + 5m questions):
2:25-2:40: Contributed Talk #1 (10min talk + 5m questions):
2:40-2:55: Contributed Talk #2 (10min talk + 5m questions):
3- 3:30 Break
3:30-3:55: Invited Talk #3 (20min talk + 5m questions):
3:55-4:10: Contributed Talk #3 (10min talk + 5m questions):
4:10-4:25: Contributed Talk #4 (10min talk + 5m questions):
4:25-4:50: Invited Talk #4 (20min talk + 5m questions):
We are soliciting abstracts (up to 1 page in length) for the Workshop on Federated KBs & the Open Knowledge Network at the Automated Knowledge Base Construction (AKBC) conference. Selected abstracts will be given short oral presentations at the workshop. Abstracts may include new unpublished work, previously published results, work in progress, datasets, open problems, etc. Submissions are non-anonymous. Relevant topics include:
James P. McCusker, Sabbir M. Rashid, Nkechinyere Agu, Neha Keshan, Deborah L. McGuinness. "A Provenance-Aware Knowledge Graph Framework for Open Knowledge Network Settings."
Jay Pujara, Louiqa Raschid, Gerard Hoberg, Gordon Phillips, Craig Knoblock. "Enterprise OKN: A Federated Knowledge Graph for Financial Data"
Varun R Embar, Jay Pujara, Lise Getoor. "Collective Alignment of Large-scale Ontologies"
Steven Haussmann, Oshani Seneviratne, Yu Chen, Yarden Ne'eman, James Codella, Ching-Hua Chen, Deborah L. McGuinness, Mohammed J. Zaki "FoodKG: A Semantics-Driven Knowledge Graph for Food Recommendation."
Please contact the organizers above for general questions.
Webpage or submission form specific questions/comments: Nicholas Monath (nmonath@cs.umass.edu)