Knowledge Base Population: Successful Approaches and Challenges

By Heng Ji, Ralph Grishman. 2011

399 citations recorded by the source corpus.

Subfields: Information Retrieval & Search

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Information Retrieval & Search

Citations in the corpus

Long Short-Term Memory-Networks for Machine Reading (2016, background): “Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pages 551–561, Austin, Texas, November 1-5, 2016. c©2016 Association for Computational Linguistics”
SciTaiL: A Textual Entailment Dataset from Science Question Answering (2018, methodology): “Standard relation identification models (Ji and Grishman 2011) assume the entity spans are provided as inputs to the model.”
Improving Information Extraction by Acquiring External Evidence with Reinforcement Learning (2016, background): “The decision d on the newly extracted values can be one of the following: (1) accept a specific entity’s value, (2) accept all entity values, (3) reject all values.”
Automated Fact Checking: Task Formulations, Methods and Future Directions (2018, background): “For example, many approaches rely on textual entailment/natural language inference (Dagan et al., 2009), while others rely on knowledge base construction techniques (Ji and Grishman, 2011).”
Multi-step Entity-centric Information Retrieval for Multi-Hop Question Answering (2019, positive): “Following the standard approach of using mention text and hyper-link information (Cucerzan, 2007; Ji and Grishman, 2011), we create a mapping (alias ta-”
Collaborative Knowledge Base Embedding for Recommender Systems (2016, methodology): “To reduce the name conﬂict issue (e.g., instead of the book, “Harry Potter” might actually mean the related movie), we extracted user’s book interests by combining un-supervised similarity computation with supervised classiﬁcation by following [13].”
TACRED Revisited: A Thorough Evaluation of the TACRED Relation Extraction Task (2020, neutral): “This makes RE a key part of many information extraction systems, and its performance determines the quality of extracted facts for knowledge base population (Ji and Grishman, 2011), or the quality of answers in question answering systems (Xu et al.”
Knowledge vault: a web-scale approach to probabilistic knowledge fusion (2014, methodology): “Entity types and predicates come from a fixed ontology, which is similar to that used in other systems, such as YAGO [39], NELL [8], DeepDive [32], and various systems participating in the TAC-KBP slot-filling competition [22].”
A Constrained Latent Variable Model for Coreference Resolution (2013, background): “…proper name or a named entity mention are more important for information extraction tasks like Wikiﬁcation (Mihalcea and Csomai, 2007; Ratinov et al., 2011), cross-document coreference resolution (Bagga and Baldwin, 1998), and entity linking and knowledge based population (Ji and Grishman, 2011).”
Mention and Entity Description Co-Attention for Entity Disambiguation (2018, background): “Among all these aspects, it has been widely accepted that semantic similarity between the context of an entity mention and its target entity candidates is the key to resolve the ambiguity (Bunescu and Pasca 2006; Cucerzan 2007; Ji and Grishman 2011; Shen, Wang, and Han 2015).”
Knowledge base completion via search-based question answering (2014, methodology): “The standard way to ﬁll in missing facts in a knowledge base is to process a large number of documents in batch mode, and then to perform named-entity disambiguation followed by relation extraction (see, e.g., Ji and Grishman [8] for a recent review).”
Explaining Trained Neural Networks with Semantic Web Technologies: First Steps (2017, background): “…need to incorporate more complex and more comprehensive background knowledge, and if readily available structured knowledge turns out to be insuﬃcient, then we foresee using state of the art knowledge graph generation and ontology learning methods [13,19] to obtain suitable background knowledge.”
Unified Structure Generation for Universal Information Extraction (2022, background): “IE tasks are highly diversiﬁed due to its varying targets (entity, relation, event, sentiment, etc.), heterogeneous structures (spans, triplets, records, etc.), and demand-speciﬁc schemas (Grishman and Sund-heim, 1996; Mitchell et al., 2005; Ji and Grishman, 2011).”