Transcript knowledge technology

Artificial Intelligence
Tibor Bosse
FACULTY OF SCIENCES
Overview
• Master Specialisations
• General facts
• Duration, Credit Points, …
• Graduation project
• Job perspectives
• Financial Issues
Overview
• Master Specialisations
• General facts
• Duration, Credit Points, …
• Graduation project
• Job perspectives
• Financial Issues
Specialisations
• Cognitive Science
• Human Ambience (started 2010/11)
• Computational Intelligence & Selforganisation
• Technical Artificial Intelligence
• Knowledge Technology & Intelligent Internet Applications
• International Semantic Web Master
Cognitive Science
(Co-operation between FPP en FEW)
What is Cognitive Science?
Two main goals:
• AI  Psy
(Use AI-techniques to study cognitive
processes)
• Psy  AI
(Use cognitive processes to make computers
more intelligent)
AI  Psychology
The approach (roughly):
1. Starting point: Psychological theory
2. Formalise as a Computer model
3. Make predictions based on this model
4. Compare predictions with reality
(empirical data)
5. Difference: Improve theory
AI  Psychology: Master Project
• Master Project: Rianne van Lambalgen
• Supervisors: Sander Los (Psy) &
Tibor Bosse (AI)
• Result: computer model for conditioning
– Kind of associative learning
– Example: reaction to starting signal
(“ready…go”)
AI  Psychology: Master Project
What is Cognitive Science?
Two main goals:
• AI  Psy
(Use AI-techniques to study cognitive
processes)
• Psy  AI
(Use cognitive processes to make computers
more intelligent)
Psychology  AI
Approach (roughly):
1. Inspiration: Psychological theory
2. Formalise as a computer model
3. Integrate model into system
– To exhibit human behaviour
(eg Computer person)
– To provide knowledge about humans
(eg Personal Assistant)
4. Evaluate the system’s behaviour
Psychology  AI: Master Project
• Master Project: Jeremy Soumokil
• Supervisor: Mark Hoogendoorn
• Goal: build an opponent for an online game,
who anticipates players’ moves.
• Method: using the concept “Theory of Mind”
– Reason about what the player is thinking
– Predict actions of the player, and anticipate on
this
Psychology  AI: Master Project
• Screenshot
Psychology  AI: Master Project
• Game setting
Psychology  AI: Master Project
• BDI model
Human Ambience
(started 2010-2011)
What is Human Ambience?
• New Master program
• Based on the idea of ‘Ambient Intelligence’
• Develop an Intelligent Environment, which
supports the human in his/her daily life
• Intelligent Car
• Intelligent House
• Intelligent Workspace
• Intelligent Phone
• …
Human Ambience - Courses
– 30 ECTS specialisation in discipline or topic
•
•
•
•
•
Criminology
Health
Mental functioning
Movement sciences
Social functioning
– Specific Human Ambience courses:
• HA Innovation
• Model-Based Intelligent Environments
Computational Intelligence and
Self Organisation
Ziezo!
What is CISO?
• Two main themes:
• Computational Intelligence
• Evolutionary Computing
• Data Mining Techniques
• Self-organising distributed systems
• Advanced Selforganisation
• Model and Analyse these systems
Inspiration – Bird Flocking
Different Profiles
• Biology
– Dynamic Energy Budgets, Evolutionary genetics,
Evolutionary Biology
• Bio-informatics
– Algorithms for Genomes, Bioinformatics data
analysis and tools, Intercellular Networks,
Computational Genomics and Proteomics
• Computer systems
– Distributed Systems, Cluster and Grid Computing,
Distributed Algorithms
• Knowledge representation
– Intelligent Web Applications, Ontology Engineering
CISO: Example Master Project
Master Project: Self-organised train dispatching
Supervisor: Martijn Schut
Organisation: Nederlandse Spoorwegen (NS)
Goal: Reducing secondary delay
Problem: too complex to plan
Solution: self-organisation, modelling...
Technical Artificial Intelligence
What is Technical AI?
• Only for students with Ba Computer Science!
• AI: trying to realise human intelligence on a
computer
• AI: different perspectives
– cognitive  psychology
– theoretical  logic
– technical  computer science
Goals TAI
• Getting familiar with developments within AI and
CS
• Learning how to make Intelligent Applications
• Design eg.
• A knowledge based system
• A multi-agent system
• A learning system …
TAI: Example Master Project
• Master Project: Sven Stam
• Supervisor: Tibor Bosse
• Goal: design ‘norm enforcing’ agent to
prevent bullying in virtual societies
• Method:
• agent-based modelling
• ‘follow’ detection module
• language processing module
• reputation maintainance module
TAI: Example Master project
• Screenshot
Knowledge Technology and
Intelligent Internet Applications
(KTIIA)
What is KTIIA?
The programme focuses on
• knowledge technology:
the technical aspects of designing knowledge systems
• knowledge management:
organisational aspects of knowledge creation, distribution,
application and maintenance
• applications of techniques from knowledge technology and
knowledge management to the Internet.
Example projects:
an academic hospital:
modelling of medical protocols for illnesses of infants
an Internet company:
recommendations about products on the Web
an accountancy checking for:
fraud detection
a personal start-up:
knowledge-based advisor for stopping with smoking
Dutch search engine company:
Semantic Web search engine
Semantic Web
WEB as a big knowledge base:
“Semantic Web”
Techniques from AI are crucial for this
new generation of the Web.
KTIIA: Example Master Project
• Project: Jacopo Urbani
• Supervisor: Frank van Harmelen
• Goal: reasoning over billions of information
units on the web of data
• Tool: using cluster of 100s of machines
– Split data
– Do reasoning on parts of the data
– Shuffle data around...
• Result: fastest existing reasoner with webdata
KTIIA: Example Master Project
KTIIA: Example Master Project
European Academy for SemanticWeb Education (EASE)
European-level Master of Science
(MSc)
Wouldn’t this be nice? (I)
Shared Master = Double Degree
Wouldn’t this be nice? (II)
Innsbruck
Amsterdam
Karlsruhe
Bolzano
Lisboa
Trento
Madrid
See other countries?
I don’t mean …
Innsbruck
Amsterdam
Karlsruhe
Bolzano
Lisboa
Trento
Madrid
Learn from the “Think Tanks” in SW
Innsbruck
Amsterdam
Karlsruhe
Bolzano
Lisboa
Trento
Madrid
Get a more complete picture
Curriculum of the Shared Master
Amsterdam
Basic Modules
Advanced
Modules
Trento
Advanced
Modules
Advanced
Modules
Bolzano
Master Project
Madrid
Advanced
Modules
Karlsruhe
Amsterdam
Specialisations
• Cognitive Science
• Human Ambience (started 2010/11)
• Computational Intelligence & Selforganisation
• Technical Artificial Intelligence
• Knowledge Technology & Intelligent Internet Applications
• International Semantic Web Master
Overview
• Master Specialisations
• General facts
• Duration, Credit Points, …
• Graduation project
• Job perspectives
• Financial Issues
Overview
• Master Specialisations
• General facts
• Duration, Credit Points, …
• Graduation project
• Job perspectives
• Financial Issues
General Facts
•
•
•
•
2 years
120 ects: 30 ects - for a graduation project
Optional space depends on variant 25-45 ects
Many other departments involved (Economy, Law,
Linguistics, Psychology, Biology, Philosophy, Social
Sciences)
General Facts
• For each variant mandatory (24 ects):
– Knowledge management and Modelling
– Evolutionary Computing
– Behaviour Dynamics
– Qualitative research Methods
– Scientific Writing in English
(+ graduation project 30 ects)
Graduation project
•
•
•
•
Last 6 months – Master Project
30 credit points
Internal or External
End product: Master Thesis
Job perspectives
• Examples:
–
–
–
–
–
–
–
–
University
Development of intelligent websites
Management of knowledge within organisations
Development of virtual agents in (serious) games
Consultancy on different aspects of planning
Various data-mining applications
Development of support systems for elderly
Development of support systems for medical patients
…
Own Company
• Both Companies and Research
Summary
This is more information than you can grasp
Go to:
http://www.few.vu.nl/onderwijs/masters/ai/
Overview
• Master Specialisations
• General facts
• Duration, Credit Points, …
• Graduation project
• Job perspectives
• Financial Issues
Overview
• Master Specialisations
• General facts
• Duration, Credit Points, …
• Graduation project
• Job perspectives
• Financial Issues
Pre-master: tuition fee and study duration
• In 2011-2012, students can follow a Pre-master by paying a fee that
is equal to the ‘wettelijk tarief’: € 1732.
• This pre-master year does not count for the ‘nominal study duration’.
• From 2011-2012 the rules will change. As soon as we know more,
you will find this on www.vu.nl/collegegeld. Advice: finish your Premaster in ’11/’12!
• Plans: students who study longer than the nominal study duration,
will pay an increased tuition fee.
• Nominal study duration for HBO students = HBO + 1 additional year
and Master + 1 additional year
Master: tuition fee and study duration
• Any student who wants to do a second Master, will have to pay
‘instellingscollegegeld’ (= € 12.000 per year)
• If you are doing two Masters simultaneously, then this will NOT hold,
as long as both Masters are finished before September 2013.
• Plans: students who study longer than the nominal study duration,
will pay an increased tuition fee (about € 3000 extra).
• Nominal study duration = standard duration of Bachelor + 1 year
and standard duration of Master + 1 year
• The ‘wettelijk tarief’ is € 1732.
Questions and Up-to-date Information
• Stand of the Student Administration
• www.vu.nl/collegegeld
• [email protected]