Case-based Reasoning in Natural Language Processing: Word2vec VS fastText

Kareem Amin, George Lancaster, Stelios Kapetanakis, Klaus-Dieter Althoff, Andreas Dengel, Miltos Petridis

In: Proceedings of the 23rd UK Workshop on Case-Based Reasoning. UK Workshop on Case-Based Reasoning (UKCBR-2018) befindet sich SGAI International Conference on Artificial Intelligence December 11-13 Cambridge United Kingdom School of Computing, Engineering and Mathematics, University of Brighton, UK 2018.


Businesses can benefi t greatly from analysing their document assets. These can vary greatly from plain text messages across customer support tickets to complex message exchanges and workflow logs within countless business transactions. Decoding text-based domain knowledge can be a challenging task due to the need for a comprehensive representation and evaluation of the business process ontology, activities, rules and paths. To provide an adequate process coverage, significant time and monetary resources should be invested as well as a high maintenance portfolio, especially for large processes and environments that change dynamically. This work investigates a novel natural language processing path which combines Case-based Reasoning and Deep Neural Networks. Our aim is to minimize the effort from domain experts while extracting domain knowledge from rich text, containing domain abbreviations, grammatically incorrect text and mixed language. Our proposed approach seems promising and a possible future direction in the industry.

Weitere Links

UKCBR-2018_paper_5.pdf (pdf, 274 KB )

Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence