Information Storage Analysis & Retrieval group www.rmit.edu.au/compsci/ infostorage
Jan 16, 2016
Information StorageAnalysis & Retrieval group
www.rmit.edu.au/compsci/infostorage
Research Focus•Web search and text information retrieval
•Data mining & machine learning
•XML & Image/Video search
•Music retrieval
•Search effectiveness
•Efficiency
RMIT University©2011 CS&IT - ISAR 2
Applications•Our in-house search engine is Zettair
–The fastest open source search engine in the world…
–…and one of the most highly effective.
•Organizing international evaluation campaigns on search of–web services–Wikipedia–web pages
RMIT University©2011 CS&IT - ISAR 3
Research staff & collaborations
•Research staff–Shane Culpepper, Simon Puglisi, Mark Sanderson, Falk Scholer, Jamie Thom, Sandra Uitdenbogerd, Jenny Zhang
•Collaborations–Companies
–Sensis, Viocorp, Funnelback, Circus Oz
–Academia–UMass Amherst, QUT, Macquarie University, University of Chile, University of Sheffield
RMIT University©2011 CS&IT - ISAR 4
RMIT University©2011 CS&IT - ISAR 5
Data Mining
sentiment analysis
bioinformatics
text mining
machine learning
–Efficient pattern discovery algorithms–Effective and novel learning models
Jenny Zhang
Algorithms for Massive Data
RMIT University©2011 CS&IT - ISAR 6
Research Strengths:
Space Efficient Data Structures Data Compression Text Processing and Indexing Natural Language Processing Distributed / Parallel Programming
Shane Culpepper
Possible Student Projects: Algorithms for Real-time Search Machine Driven Search Data Compression Algorithms Data Streaming Algorithms Persistent and Parallel Data Structures
Language Independent Text Indexing IR Applications of Self-Indexes Applying NLP in Information Retrieval
Applications of Metadata and Multimedia Retrieval
•Accounting for SustainabilityRepresenting and querying knowledge about sustainability indicators using XML, RDF, OWL, SPARQL
•The Circus Oz Living ArchiveCombination of– content based image and video retrieval
– tagging of video
RMIT University©2011 CS&IT - ISAR 7
James Thom
Social Information Search
Mark Sanderson
•Search interfaces and result presentation
•Measurement of performance
•User-based evaluation – what is a “useful” answer?
•Effective summarisation of documents
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Precision
Recall
Search Effectiveness
RMIT University©2011 CS&IT - ISAR 9
Mark Sanderson, Audrey Tam, Falk Scholer