CALL FOR PARTICIPATION
Large, publicly available knowledge graphs (KGs) have given rise to many successful improvements on traditional information retrieval tasks such as ad hoc document retrieval. This workshop focuses on the end-to-end utilization of knowledge graphs and semantics in text retrieval and other IR-related applications. Its scope covers the acquisition, the alignment, and the utilization of knowledge graphs and semantic resources for the purpose of optimizing end-to-end performance of a system that responds to a user’s information need. Examples of such technologies and applications include entity ranking, entity linking, entity-based retrieval models, entity recommendation, document filtering, knowledge graph population, and more. This workshop features discussions and presentations on innovative ideas for new methods, suggestions for shared tasks and benchmarks, position papers, as well as reports on practical experiences with KG technology both from academia and industry.
Topics of this workshop include but are not limited to:
- Acquisition for IR applications
- knowledge graph population for IR
- domain/task-specific knowledge graph construction
- knowledge representation for IR
- query-time knowledge extraction
- Alignment with and for queries
- entity linking for short queries and documents
- relation extraction for query and document modeling
- information integration and ontology matching
- entity search
- subgraph ranking
- Utilization for text retrieval
- ad hoc document retrieval
- passage retrieval
- question answering
- dialogue systems
- event tracking and summarization
- complex answer retrieval
This workshop features keynotes, a discussion panel, as well as poster and oral presentations.
- Keynotes will be presented by invited speakers from industry and/or academia who are leading experts in relevant topics.
- The panel discussion will host discussions about current progress, processes, challenges, joint efforts, and potential research directions in KG4IR.
- Oral presentations will feature presentations from accepted papers with the option of also presenting them as posters; those with highest potential to spark discussions will be favored in the selection process.