Workshop on Personal Knowledge Graphs

The concept of personal knowledge graphs (PKGs) has been around for a while, in recognition of the need to represent structured information about entities that are personally related to a user. However, several open questions remain:

• Definition — The notion of a personal knowledge graph has been established loosely, as a resource of structured information about entities personally related to its user. This definition needs crystallization: What is personal knowledge and how is it represented in a PKG? What differentiates a PKG from general KGs, how are they related? How can PKGs benefit from information stored in general KGs and how is the benefit realised? How is work on PKG related to work in areas such as commonsense KGs and entity/event-centric understanding?
• PKG construction/population — What are the potential data sources (textual, visual, geolocation, etc.)? How to mitigate the ‘dual use’ of automated technology for PKG construction/population, since it can be used to extract and exploit personal knowledge about others?
• PKG utilities — What novel application scenarios would PKGs enable and what role does/can novel techniques such as semantic technologies and knowledge modeling play in this respect? How do PKG compare to existing solutions to these applications?
• Practical realization — Where would PKGs be stored and how would these interact with a range of external services, while considering access control as well as privacy concerns of users?

Topics of Interest

The goal of the PKG workshop is to create a forum for researchers and practitioners from diverse areas to present and discuss methods, tools, techniques, and experiences related to the construction and use of personal knowledge graphs, identify open questions, and create a shared research agenda.

We invite submissions of regular papers, position papers, demonstrators as well as encore talks (featuring already published work).

Topics of interest include, but are not limited to:

• Modeling personal knowledge (ontologies and knowledge representation)
• Populating and maintaining personal knowledge graphs (natural language processing, entity and relation extraction, information integration)
• Applications of personal knowledge graphs (including but not limited to intelligent personal assistants, personal information management systems, health information systems, recommender systems)
• Evaluation (evaluation measures and methodology, benchmark construction)
• Systems and toolkits

Submission Guidelines

We welcome submissions of regular papers (4-6 pages) that present original technical, theoretical, or experimental contributions; position papers (2-4 pages) that explore controversial, risk-taking or nascent ideas that have the potential to spark debate and discussion at the workshop; demonstrator papers (max 4 pages) that present first-hand experience with research prototypes or operational systems. Submissions are single-blind and should use a double-column CEUR-WS proceedings format.

In addition to the paper contributions, we also invite submissions of encore talks, to present work that has already been published in a leading conference or journal, but is relevant to the topics of this workshop.  We require a copy of the paper along with a brief (max 200 words) statement of relevance.

Submission site: https://easychair.org/my/conference?conf=pkg2021

Workshop Organizers

• Krisztian Balog, University of Stavanger, Norway
• Paramita Mirza, Max Planck Institute for Informatics, Germany
• Martin G. Skjæveland, University of Oslo, Norway
• Zhilin Wang, University of Washington, USA

Feel free to contact us at pkg2021@easychair.org!

Registration

Workshop registration will be through the AKBC 2021 conference registration.

Program Committee

• Faegheh Hasibi, Radboud University, Netherlands
• Kjetil Kjernsmo, Inrupt, Norway
• Filip Radlinski, Google, London
• Anna Tigunova, Max Planck Institute for Informatics, Germany
• Ruben Verborgh, Ghent University, Belgium
• Gerhard Weikum, Max Planck Institute for Informatics, Germany
• Charles Welch, University of Michigan, USA

Keynote

Speaker: Cathal Gurrin (Dublin City University)

“Lifelogs as Personal Data Information Systems“

Abstract:
As new forms of personal data become prevalent, the challenge for information scientists is in facilitating the individual to make use of this data in daily life. Lifelogs, which lie at the more extreme end of the personal data spectrum, are large multimodal archives of an individual’s activities captured using various software and hardware sensors over long time periods. In event years, we have seen the first collaborative benchmarking activities in this domain, which are facilitating the research community to progress research into lifelog information systems in an open and collaborative process. In this talk I will motivate researching into lifelog information systems and highlight the progress being made in lifelog information system’s management through benchmarking activities such as the annual Lifelog Search Challenge. Such techniques include conventional inverted indexes, embedded spaces, self-organising concept hierarchies and lifelog knowledge graphs. The talk will conclude with a call to action for the community and suggestions of the most impactful research activities.

Accepted Contributions

• [regular – R1] Personal Health Knowledge Graph for Clinically Relevant Diet Recommendations (Oshani Seneviratne, Jon Harris, Ching-Hua Chen and Deborah McGuinness)
• [encore – E1] Data Augmentation for Personal Knowledge Base Population (Lingraj S Vannur, Balaji Ganesan, Lokesh Nagalapatti, Hima Patel, and MN Thippeswamy)
• [encore – E2] Extracting Personal Information from Conversations (Anna Tigunova, Andrew Yates, Paramita Mirza, and Gerhard Weikum)
• [encore – E3] Bias in Conversational Search: The Double-Edged Sword of the Personalized Knowledge Graph (Emma J. Gerritse, Faegheh Hasibi, and Arjen P. de Vries)
• [encore – E4] PROV4ITDaTa: Transparent and direct transfer of personal data to personal stores (Gertjan De Mulder, Ben De Meester, Pieter Heyvaert, Ruben Taelman, Ruben Verborgh, and Anastasia Dimou)
• [encore – E5] A Knowledge Base for Personal Information Management (David Montoya, Thomas Pellissier Tanon, Serge Abiteboul, Pierre Senellart, and Fabian M. Suchanek)
• [encore – E6] Dialogue-Based Relation Extraction (Dian Yu, Kai Sun, Claire Cardie, and Dong Yu)
• [encore – E7] Toward Activity Discovery in the Personal Web (Tara Safavi, Adam Fourney, Robert Sim, Marcin Juraszek, Shane Williams, Ned Friend, Danai Koutra, and Paul N. Bennett)
• [encore – E8] You Get What You Chat: Using Conversations to Personalize Search-Based Recommendations (Ghazaleh H. Torbati, Andrew Yates, and Gerhard Weikum)
• [encore – E9] YourDigitalSelf: A Personal Digital Trace Integration Tool (Varvara Kalokyri, Alexander Borgida, and Amélie Marian)

Workshop Schedule

	Time (PST)	Time (CET)	Title
Session 1	08:00 – 08:10	17:00 – 17:10	Intro & Welcome
	08:10 – 08:50	17:10 – 17:50	Keynote
	08:50 – 09:00	17:50 – 18:00	Summary from CASK workshop
	09:00 – 09:30	18:00 – 18:30	Accepted Contributions: R1, E1 (12+3 min per talk+QA)
break	09:30 – 10:00	18:30 – 19:00
Session 2	10:00 – 11:30	19:00 – 20:30	Accepted Contributions: E2–7 (12+3 min per talk+QA)
break	11:30 – 12:00	20:30 – 21:00
Session 3	12:00 – 12:30	21:00 – 21:30	Accepted Contributions: E8–9 (12+3 min per talk+QA)
	12:30 – 12:35	21:30 – 21:35	Forming breakout groups
	12:35 – 13:10	21:35 – 22:10	Breakout group discussion
	13:10 – 13:30	22:10 – 22:30	Reporting back from breakout groups