IEEE ACT CS Seminar 1 - Entity Resolution and Social Media Analysis
Seminar with two talks: Entity Resolution and Social Media Analysis, Thursday 4 April
Talk one: Certus - An Effective Entity Resolution Approach with Graph Differential Dependencies (GDDs)
Speaker: Dr Jixue Liu, University of South Australia
Talk two: Using social media to predict, detect and summarise protest events
Speaker: A/Prof. Lin Liu, University of South Australia
Date and time: 3:45-4:45pm, Thursday 4 April 2019
Location: CSIRO Discovery Centre, Black Mountain (on North Science Road)
Room: Discovery Theatre (on right-hand side at entry of Discovery Centre)
Date and Time
Location
Hosts
Registration
-
Add Event to Calendar
- Black Mountain
- Canberra, Australian Capital Territory
- Australia
- Building: CSIRO Discovery Centre
- Contact Event Host
- Co-sponsored by Yanchang Zhao
Speakers
Dr. Jixue Liu of University of South Australia
Certus: An Effective Entity Resolution Approach with Graph Differential Dependencies (GDDs)
Entity resolution (ER) is the problem of accurately identifying multiple, differing, and possibly contradicting representations of unique real-world entities in data. In this work, we propose graph differential dependencies (GDDs) as an extension of the recently developed graph entity dependencies to enable approximate matching of values. Furthermore, we investigate a special discovery of GDDs for ER by designing an algorithm for generating a non-redundant set of GDDs in labelled data. Then, we develop an effective ER technique, Certus, that employs the learned GDDs for improving the accuracy of ER results. We perform extensive empirical evaluation of our proposals on five real-world ER benchmark datasets and a proprietary database, and the results show our approach significantly improves the precision of ER without considerable trade-off of recall.
Biography:
Dr Jixue Liu is a senior lecturer in the University of South Australia. He got his PhD in computer science from the University of South Australia. His research interests include integrity constraint discovery, data analytics in texts and timeseries data, entity linking, algorithmic discrimination and privacy. He has done work in XML functional dependencies and query translation, trust management on the internet, integration and transformation of data, and view maintenance.
A/Prof. Lin Liu of University of South Australia
Using social media to predict, detect and summarise protest events
Social media sites such as Twitter have been observed to be viable platforms for facilitating civil unrest related conversations as well as mobilizing people for protests. Deriving insights into protest events help to understand why, when, and how the events occur. Such insights are beneficial to law enforcement agencies, emergency services, security analysts, governments, etc. In this research, we propose to leverage the rich information propagation in the Twitter network community structure to develop models that help to derive insights into civil unrest. We have developed three models, namely Propagation Tree, SensorTree and StoryGraph, to predict, detect and summarize protest events using social media.
Biography:
Dr Lin Liu is an associate professor at the School of IT and Mathematical Sciences, University of South Australia. Her research interests include data mining, causal inference and bioinformatics. She has published over 90 journal/conference papers and 1 book. Dr Liu's research has been supported by 3 ARC discovery projects and several CRC projects. Dr Liu is a conference chair of 2019 Australasian Data Mining Conference (AusDM 2019), a PC chair AusDM 2013, 2014 and 2017, and an organiser of KDD Causal Discovery Workshop (2016-2018).
Agenda
3:45-4:15pm:
Talk one: Certus - An Effective Entity Resolution Approach with Graph Differential Dependencies (GDDs)
4:15-4:45pm
Talk two: Using social media to predict, detect and summarise protest events