Jaideep Srivastava

1
Jaideep Srivastava

• University of Minnesota
• srivasta_at_cs.umn.edu

US-China Workshop on Collaboration for Tobacco
Control Research March 27-29, 2008 Beijing, PRC
2
Scientific aim of research in informatics

• New information models
• Develop information models and metrics to capture
  the needs of novel applications
• New algorithms and data structures
• Develop computational models and algorithms to
  capture the processing needs of novel
  applications
• Accuracy and scalability
• Improve the accuracy and scalability of existing
  algorithms
• Computational tools
• Develop computational tools to make research
  applicable to researchers and practitioners

3
How can technology improve ?

• data collection
• very high resolution, ubiquitous, comprehensive
  data collection, and low cost, high capacity
  storage
• collect everything you want, store it for as long
  as you care, at practically no cost
• data management
• low cost, high capacity storage, extremely fast
  processing, high speed networks, ubiquitous
  access
• data analysis
• data mining/machine learning, scalable
  algorithms, handling high-dimensional, sparse
  datasets
• project implementation
• higher level software tools/modules, open source
  software, public datasets
• research collaboration
• networks, collaboration tools,
• network development
• easier to interact ? people-2-people n/ws easier
  to develop/maintain
• capacity building
• increased awareness and interest on the part of
  young people

4
Case examples demonstrating how your technology
addressed problems

• MINDS Minnesota Intrusion Detection System

MINDS system
Association pattern analysis
Summary of attacks
Anomaly scores
network
Detected novel attacks
Anomaly detection

Humananalyst
MINDSAT
Data capturing device
Labels
Known attack detection
Detected known attacks
Feature Extraction
Filtering

• The MINDS is being used at University of
  Minnesota and at the ARL Center for Information
  Monitoring and Protection to detect attacks and
  intrusive behavior that can not be detected using
  widely used intrusion detection systems, such as
  SNORT.
• Anomaly detection techniques are used to discover
  novel and previously unknown attacks
• Association rule based techniques are used to
  summarize and characterize suspicious network
  connections

5
Learning experiences for future efforts

• Rapid increase in the usage of the Internet, Web,
  laptops, PDAs, cell phone, etc.
• ? increased amount of data generation
  activities on line
• Increase in resolution and deployment of data
  collection probes, e.g. GPS, cell phone/PDA,
  wireless enabled laptop, RFID tags,
• ? increased ability to monitor and record
  interactions at a really fine granularity
• Dramatic increase in storage capacity and
  decrease in storage costs
• ? feasible to store all the data collected

Great opportunity to (i) advance the state of the
art in data analysis technologies, and (ii) use
them to better understand various phenomena