Computational narrative mapping for the acquisition and representation of lessons learned knowledge. (May 2018)

• Record Type: Journal Article
• Authors: Yeung, C.L.; Wang, W.M.; Cheung, C.F.; Tsui, Eric; Setchi, Rossitza; Lee, Rongbin W.B.
• Is Part Of: Engineering applications of artificial intelligence. Volume 71(2017:Nov.)
• Page Start: 190
• Page End: 209
• Publication Date: 2018-05
• Subjects: Knowledge management -- Lessons learned -- Knowledge acquisition -- Knowledge representation -- Human learning -- Computational narrative mapping; Engineering -- Data processing -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; Ingénierie -- Informatique -- Périodiques; Intelligence artificielle -- Périodiques; Systèmes experts (Informatique) -- Périodiques; Artificial intelligence; Engineering -- Data processing; Expert systems (Computer science); Periodicals; 620.00285
• DOI: 10.1016/j.engappai.2018.02.011
• Languages: English
• ISSNs: 0952-1976

Abstract

Abstract: Lessons learned knowledge is traditionally gained from trial and error or narratives describing past experiences. Learning from narratives is the preferred option to transfer lessons learned knowledge. However, learners with insufficient prior knowledge often experience difficulties in grasping the right information from narratives. This paper introduces an approach that uses narrative maps to represent lessons learned knowledge to help learners understand narratives. Since narrative mapping is a time-consuming, labor-intensive and knowledge-intensive process, the proposed approach is supported by a computational narrative mapping (CNM) method to automate the process. CNM incorporates advanced technologies, such as computational linguistics and artificial intelligence (AI), to identify and extract critical narrative elements from an unstructured, text-based narrative and organize them into a structured narrative map representation. This research uses a case study conducted in the construction industry to evaluate CNM performance in comparison with existing paragraph and concept mapping approaches. Among the results, over 90% of respondents asserted that CNM enhanced their understanding of the lessons learned. CNM's performance in identifying and extracting narrative elements was evaluated through an experiment using real-life narratives from a reminiscence study. The experiment recorded a precision and recall rate of over 75%. Highlights: Learning from narratives of past experiences is vital to transfer lessons learned knowledge. Narrative maps are used to represent lessons learned knowledge. Computational narrative mapping (CNM) is used to automate the narrative mapping process. A prototype of CNM was built and trial implemented in the construction industry. The results show that CNM performs significantly better than existing approaches.