SMART 2022 was the third edition of SMART, the subtasks for Question Answering over Knowledge Graph, which is part of the ISWC 2022 Semantic Web Challenge. It was co-located with the 21st International Semantic Web Conference (ISWC 2022)1. The first edition SMART2020 [1] was in ISWC 2020 and the second edition SMART2020 [2]. Given a question in natural language, the task of the SMART challenge is, to predict the answer type, entities and relations using a target ontology. In the current (third) edition, the SMART challenge had three tracks i.e. answer type prediction, entity linking and relation prediction. These three tasks are based on two popular KBs, one using the DBpedia ontology and the other using Wikidata. This volume contains peer-reviewed system description papers of all the systems that participated in the challenge. More details about the challenge can be found at https://smart-task.github.io/2022/. This challenge is focused on answer type prediction, entity linking and relation prediction, which play an important role in KGQA (Question Answering over Knowledge Graphs) systems. Answer Type Prediction Given a natural language question, the task is to produce a ranked list of answer types of a given target ontology. Previous such answer type classifications in literature are performed as a short-text classification task using a set of coarse-grained types, for instance, either six types [3, 4, 5, 6] or 50 types [7] with TREC QA task2. We propose a more granular answer type classification using popular Semantic Web ontologies such as DBpedia and Wikidata.
Relation Prediction Given a natural language question, the task is to identify the relation and link to the relations in KG. Depending on the number of relations in the KG, the number of relation types to be linked varies. Entity Linking Given a natural language question, the task is to identify the entity mentions and link to the entities in KG.
Table 1, Table 2 and Table 3 illustrate some examples. The participating systems can be either supervised (training data is provided) or unsupervised. The systems can utilize a wide range of approaches; from rule-based to neural approaches. Organisation
dbo:influencedBy
wdt:P737 dbo:starring, dbo:director
wdt:P161, wdt:P57 dbo:numberOfEmployees wdt:P1128 In this section, we list the people who organised and contributed to the success of this event. Challenge Chairs • Nandana Mihindukulasooriya (IBM Research AI) • Mohnish Dubey (Philips Research, Netherlands) • Alfio Gliozzo (IBM Research AI) • Jens Lehmann (Amazon) • Axel-Cyrille Ngonga Ngomo (Universität Paderborn) • Ricardo Usbeck (University of Hamburg) • Gaetano Rossiello (IBM Research AI) • Uttam Kumar (University of Bonn) • Debayan Banerjee (Universität Hamburg)
dbr:Perl
wd:Q42478 dbr:William_Shatner
wd:Q16297 dbr:IBM wd:Q37156 Challenge Programme Committee Members The challenge program committee helped to peer-review the eight system papers. Each paper received 2 or 3 reviews from the program committee members and authors took those feedback into account when preparing the camera-ready versions. The organizers would like to thank them for their valuable time. • Ahmad Alobaid (Universidad Politécnica de Madrid) • Debayan Banerjee (Universität Hamburg) • Mohnish Dubey (Philips Research, Netherlands) • Longquan Jiang (University of Hamburg) • Marcos Martinez-Galindo (IBM Research) • Nandana Mihindukulasooriya (IBM Research AI) • Cedric Möller (University of Hamburg) • Ricardo Usbeck (University of Hamburg) We would like to thank the ISWC Semantic Web Challenge chairs, Daniele DellAglio, Catia Pesquita and the whole ISWC organising committee for their invaluable support to make this event a success. We would also like to thank the challenge participants for their interest, quality of work, and informative presentations during the event which made it attractive to the ISWC audience.