Author: Li, Yiu Keung
Title: A study of computational and human models of serendipitous information seeking
Advisors: Tsui, Eric (ISE)
Lee, W. B. (ISE)
Degree: Eng.D.
Year: 2020
Subject: Hong Kong Polytechnic University -- Dissertations
Information retrieval
Human information processing
Department: Faculty of Engineering
Pages: 317 pages : color illustrations
Language: English
Abstract: This research is about serendipitous encounters in information. The first mentioned serendipitous encounter was the first story of Steve Jobs mentioned in a speech at Stanford University, which told of connecting the "dots". That was the start of the preparation of my thesis concerning about the serendipitous information seeking. The connection of the "dots" is the most important idea and the connection of this idea to serendipity is a significant contribution of this thesis using computational/algorithmic and human models. Serendipity is about finding useful deviations from expected patterns so as to provide new insights (Rond, 2005) (Foster, 2004) (Foster & Ford, 2003) (Yaqub, 2018). With an advanced technology like data/text mining (Han, et al., 2011), the opportunity of serendipity on finding hidden patterns from a large corpus of documents has never been higher. The existing literature of computational serendipity models or the like (Corneli, et al., 2014) (Rond, 2005) (Figueiredo & Campos, 2001) (McCay-Peet & Toms, 2011) (Yaqub, 2018) covers many areas of human models such as serendipity triggers, focus shifts, prepared minds, and other bridging conditions, and so forth. Data/text mining tools can work on those areas finding surprising useful insights. The serendipitous results of text mining using existing knowledge management literature and a suite of tools called Chance Discovery (CD) (Wang & Ohsawa, 2011) with some computational algorithms (e.g. KeyGraph) (Ohsawa, et al., 1998) (Wang, et al., 2013) (Zhang, et al., 2013) were examined with the human factors mentioned above for better opportunities to experience serendipity.
Latent Dirichlet Allocation (LDA) (Blei, et al., 2003) (Hoffman, et al., 2010) is a generative topic discovery algorithm (Ponweiser, 2012) which builds probabilistic models of latent topics for a corpus of document. The LDA algorithm was accidentally found to be the long-waiting tool for detecting serendipitous findings from a large volume of text. The possible ways to experience serendipity were first investigated by referring to some existing human model-related theories and frameworks of cognition, learning, linguistics, entrepreneurship, creativity, data mining, and knowledge management. Afterwards, the serendipitous nature of the LDA algorithm could be interpreted by referring and mapping its properties to the theories and frameworks to derive a far more comprehensive serendipitous information seeking model. Several evaluation sessions were conducted with around twenty participants invited to test serendipitous information findings. The test results as to whether the participants experienced serendipity were presented by using text mining technologies with some Knowledge Management literature found from random searching using Google. The important algorithms, including Keygraph, LDA/LDAvis (Sievert & Shirley, 2014), & Theme Evolutionary Graph (TEG), were evaluated to see how the serendipity experience could be identified. Based on the evaluation results, the supposition of serendipity as the connection of the "dots" serving like the "preparedness" was reviewed to explore the viability. Potential development was mapped out by recognizing the possibility of using narrative creation and scenario planning as human models to experience serendipity. The evaluation results presented in this thesis indicated the potential to experience serendipity in information seeking via computational/algorithmic approaches.
Rights: All rights reserved
Access: restricted access

Files in This Item:
File Description SizeFormat 
991022378657303411.pdfFor All Users (off-campus access for PolyU Staff & Students only)8.89 MBAdobe PDFView/Open

Copyright Undertaking

As a bona fide Library user, I declare that:

  1. I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
  2. I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
  3. I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.

By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.

Show full item record

Please use this identifier to cite or link to this item: