Norsk English

Entity lookup for Knowledge Graphs

The goal is to identify and extract entities from the Knowledge Graph that best match or relate to the input text, a crucial process for fields like information extraction, question answering systems, and semantic search [Avogadro et al 2022, Gillick et al 2019].

Contact person

Keywords: Data linking, Data enrichment, Tabular data, Knowledge Graphs (KGs)

Knowledge Graphs (KGs), with their interconnected entities and relationships, are pivotal in applications like search engines and AI. A key operation in KGs is entity lookup, which retrieves entities using specific text inputs. Given the large scale and complexity of KGs, this task can be challenging.

In information retrieval, text-based entity lookup employs natural language processing techniques like Named Entity Recognition (NER) and Term Frequency-Inverse Document Frequency (TF-IDF), and potentially advanced methods like machine learning. The goal is to identify and extract entities from the KG that best match or relate to the input text, a crucial process for fields like information extraction, question answering systems, and semantic search [Avogadro et al 2022, Gillick et al 2019].

Expanding on the existing solution at https://bitbucket.org/disco_unimib/lamapi, the aim is to optimise the data retrieval process for better performance and quicker response time. The objective includes enhancing the system's robustness against errors. To measure the enhancements, Coverage and Mean Reciprocal Rank (MRR) will be used as validation metrics.

Work to be done:

  •  Develop a proposal for a method more efficient than the current one.
  • Evaluate and present the response time of the proposed method in terms of queries per second, using a given infrastructure.
  • Validate the effectiveness of the proposed method using Coverage and Mean Reciprocal Rank (MRR) as key metrics.

References:

[Avogadro et al 2022] Avogadro, R., Cremaschi, M., D’adda, F., De Paoli, F., & Palmonari, M. (2022). LamAPI: a Comprehensive Tool for String-based Entity Retrieval with Type-base Filters. In 17th ISWC workshop on ontology matching (OM).

[Gillick et al 2019] Gillick, D., Kulkarni, S., Lansing, L., Presta, A., Baldridge, J., Ie, E., & Garcia-Olano, D. (2019). Learning dense representations for entity retrieval. arXiv preprint arXiv:1909.10506.