Name Clarifier: A Visual Analytics System for Author Name Disambiguation

The Hong Kong University of Science and Technology 
Department of Computer Science and Engineering 

Final Year Thesis Oral Presentation

Title: "Name Clarifier: A Visual Analytics System for Author Name 
        Disambiguation"

by 

Ms. WU Tongshuang, Sherry


Abstract: 

Distinguishing authors with similar names is an important prerequisite 
for various problems at individual levels, and thus has been widely 
studied recently. However, most existing mining algorithms could only 
partially solve this problem. This is because name ambiguity has very high 
variety, which could not be easily captured by a general model. 
Consequently, trickier cases are often missed. On the other hand, humans 
are often more capable of distinguishing one author from another by taking 
advantages of some expertise knowledge. Thus, we present a novel visual 
analytics system, NameClarifier, to facilitate name disambiguation by 
keeping the experts in the loop. Specifically, NameClarifier quantifies 
and visualizes the similarities between unidentified authors and the ones 
that have been successfully disambiguated using three key factors, namely 
co-authorships, publication venues and temporal information. Then, the 
disambiguation is performed iteratively: The papers that have apparent 
evidence are still automatically disambiguated, and the more ambiguous 
cases are handed to users for manual identification with visualization. 
Given visual cues, users are implicitly guided to weigh different factors 
on a case-by-case basis. Such customization distinguishes authors better 
than identical models for some extreme cases. The new disambiguation, once 
processed, would instantly serve as additional cues for all the remaining 
unidentified authors. In this way, we open up the black box in traditional 
name disambiguation processes, and help intuitively and comprehensively 
reason why the corresponding classifications should hold. Case studies and 
a user study are conducted to demonstrate the usefulness of NameClarifier 
in the accuracy and the reasoning of name disambiguation.

  
Date                 : 27 April 2016 (Wednesday)

Time                 : 2:30pm to 3:30pm

Venue                : Room 5562 (lift 27/28)

Advisors             : Prof. Qiang YANG
                     : Prof. Huamin QU

2nd Reader           : Dr. Qiong LUO