Yesterday Today, and Tomorrow of AI applications

Download Report

Transcript Yesterday Today, and Tomorrow of AI applications

Brain Amplifier for Holistic Knowledge
Management using New Generation of AI
Dr. Eunika MERCIER-LAURENT
VP AFIA
MODEME UMR 5055- Research Center
IAE Université Lyon 3, France
ECCAI ASTI
AI was born from dreams
It is still what we need to innovate
•~3000BC first known expert system
•13th C Ramón Lull invented the Zairja, the first device that systematically tried to generate
ideas by mechanical means
•17th first computer Pascal and Leibnitz
•1921 robot (Karel Capek)
•1945 ENIAC Electronic Numerical Integrator and Calculator
•1945 – 1956 cybernetics, neural nets – learning (Hoebbs)
•1950 Turing test to measure machine intelligence
•1956 Logic theorist first AI pg A. Newell, H. Simon and JC Shaw
•1956 AI was born, the founders was supposed to understand human intelligence and ready to
put it into machine
•Checkers-playing pg, GPS, NL (Chomsky), perceptron, Dendral, automatic translation…
•1970 machine learning, Prolog
•1975 XCON first commercial expert system, speech recognition
•1980 first commercial tools, beginning of industrial expert systems, fuzzy logic, genetic
algorithms
• 1985 constraint programming
•1990 CBR, conceptual knowledge modeling, ontology…
More on http://www.crl.ucsd.edu/~elman/Courses/cog202/Papers/ai-history.html
and on http://emlkmi.free.fr/tiki-index.php?page=HistoireIa
ICAI'06
Eunika Mercier-Laurent
2
AI in France
– Prof. Jacques Pitrat, Paris 6
– Natural language processing: Prolog Marseille 1970,
automatic translation, NL interface to DB
– INRIA NL fuzzy information retrieval from DB (80)
workstation for collaborative work with AI inside (81)
– Groupe Bull :
• Research Center 1981 (NL to DB, KOOL, Prolog)
• ECRC 1984 (constraint prog, Prolog machine, deductive
DB
• CEDIAG 1985 : KOOL, Charme, KADS, EDEN,
applications
• Global approach (91), Organizational Memory, ontology
1994
– ECCAI 82, AFIA 89, IJCAI 93,141 research teams, 37 AI
comp Ilog, Cosytec, Kaidara…
– Conferences RFIA (since1980) & Plate-forme (since 1999)
including applications
– AI in national and European Programs
ICAI'06
Eunika Mercier-Laurent
3
CEDIAG main applications
ALPIN Expert system for Medical insurance with natural language module for automatic
processing of medical reports;
NOEMIE Configuration system for Bull computers
Diagnosis and help desk for customer support,
KRONES configuration support system and diagnosis for bottle-washing engines
Danish Customs’ decision support system for interpretation of ECC regulations. A similar system
was later developed for Argentina Customs.
ARAMIS-GM French national Guards Missions planning system (resource allocation, crisis
situation management). The first hybrid system was composed of database natural language
retrieval, expert system and constraint programming techniques.
RAMSES Security of the Winter Olympic games Albertville 1992, in which we have reused our
experience from ARAMIS-GM development.
SACHEM, decision support system for blast furnaces, the largest European AI application for
Sollac (Groupe Arcelor).
Knowledge acquisition from telemetric data for Formule1 racing cars, reusing prior experience.
Computer network diagnosis,
Optimized keys designing,
Scheduling, time-tables for colleges, universities and engineering schools,
Planning and resources allocation for orange picking and optimizing juice production….
ICAI'06
Eunika Mercier-Laurent
4
What we learnt ?
Knowledge is relative to human
Human is complex
Human have to deal with complex systems
ICAI'06
Eunika Mercier-Laurent
5
Context of AI since 1996
• Name : AI or computational intelligence ?
• AI promised to much ?
• AI inside is not AI ?
Decision support systems for technical and medical
diagnosis, help desk, maintenance, scheduling,
optimization, risk analysis, process control, traffic
control
design, advisory systems, software..
Solutions are mainly software, we can do better
ICAI'06
Eunika Mercier-Laurent
6
Opportunity for AI
• Internet : Information overload, lot of data
e-services, e-business, e-learning, e-government, m-ware,
content design and search …without AI,
• Knowledge Management is about Knowledge ! (k transfert,
sharing, finding, learning, innovation)
lack of feedback from AI applications and ..reinventing the
wheel
• Mondialization : complex problem to solve, cultures, access
to collective K, RT translation
ICT : What is missing ? I
ICAI'06
AI = different thinking
Eunika Mercier-Laurent
7
AI Today
Integration of existing techniques (but nothing
really new since 20 years)
–
–
–
–
–
Knowledge Discovery from multimedia documents
Semantic web (ontology, NLP, MAS, ANN)
Collective intelligence for MAS
Artificial life (understanding brain is to difficult ?)
Hybrid solutions (soft) for complex problems (text
mining, automatic indexing and retrieval..)
– Intuitive HMI
– Computer graphics, VR, immersion, simulation
– Co-design
Classic and AI
ICAI'06
Eunika Mercier-Laurent
8
New generation of AI (symbolic and robotics)
ICAI'06
Eunika Mercier-Laurent
9
Future of AI
Welcome to my dreams
ICAI'06
Eunika Mercier-Laurent
10
C3PO and R2D2
from Star Wars
Some inspirations
• Leonardo da Vinci, Jules Verne, Isaac Asimow,
Lew Bobrow, Stanislaw Lem…
• AI, Minority report, Star Wars
• More human robot ? Johny 5 from Short Circuit
• More powerful human ? Extension of biological
capability as Jake Foley…
ICAI'06
Eunika Mercier-Laurent
11
Five Generations of Management Styles
Adapted from Innovation Strategy for The Knowledge Economy Debra M. Amidon 1997 BH
1st Technology
as the Asset
2nd Project
as the Asset
3rd Enterprise
as the Asset
4th Customer 5th Knowledge
as the Asset
as the Asset
Core
Strategy
• R&D in Isolation
• Link to
Business
• Technology/
Business
Integration
• Integration
With Customer
R&D
• Collaborative
Innovation
System
Change
Factors
• Unpredictable
Serendipity
• Interdependence
• Systematic R&D
Management
• Accelerated
Discontinuous
Global Change
• Kaleidoscopic
Dynamics
• R&D as
Overhead
• Cost-Sharing
• Balancing
Risk/Reward
• ‘Productivity
Paradox’
• Intellectual
Capacity/
Impact
Structure
• Hierarchical;
FunctionallyDriven
• Matrix
• Distributed
Coordination
• Multi-Dimensional
‘Communities of
Practice’
• Symbiotic
Networks
People
• We/They
Competition
• Proactive
Cooperation
• Structured
Collaboration
• Focus on
Values and
Capability
• Knowledge
Cultivators
Process
• Minimal
Communication
• Project-toProject Basis
• Purposeful
R&D/Portfolio
• Feedback Loops • Cross-Boundary
Learning and
and ‘information
Knowledge Flow
persistence’
• Data-Based
• InformationBased
• IT as a
Competitive
Weapon
Performance
Technology
ICAI'06
• Embryonic
Customer Retention Eunika Mercier-Laurent
Customer Satisfaction
• Intelligent
Knowledge
Processors
12
Customer Success
Technology replaced person-to-person interactions with
person-to-machine…
Technology requires the synergy of individuals, machines and
social organizations and depends profoundly both on an
understanding of nature - on science - and on the capability to
design.
Virtually every human activity - agriculture, commerce,
education, health care, warfare, industry and more - depends
directly or indirectly on our interactions as individuals with
society and machines.
George Bugliarello
ICAI'06
Eunika Mercier-Laurent
13
My vision
In my vision, connected human knowledge cultivators
work in perfect synergy with the artificial
knowledge processors.
They learn from each other.
Computers help people by performing the tasks
difficult or impossible for human to do…
in the world where the biological, social and machine
components are well balanced, are sustainable
indefinitely without destroying the environment,
and enhance the human condition.
ICAI'06
Eunika Mercier-Laurent
14
What kind of machines ?
Spirit
Aibo
Electrolux
Automower
ICAI'06
Eunika Mercier-Laurent
15
Innovate for what ?
Dreams
Problems & Needs
(market)
Power of (imagination & intelligence))
methods
techniques
solutions
ICAI'06
Products and
services
Eunika Mercier-Laurent
16
Problems and needs
Today problems are complex
challenge : find balanced solution
Needs
–
–
–
–
–
–
–
–
–
ICAI'06
Safety of persons and systems (cryptography, identity management, intrusion
detection, security at home, security of Information Systems)
Health, k of our body and k about how to care it, human « spare parts »
Sustainability
• Agriculture (ancestral knowledge rediscovering instead of inventing new
pesticides and artificial fertilizers) RICH
Intelligent services on line FAQ…
social (loneliness, handicapped, unemployed),
H & C Learning (3W access to K), interactive, with RV, collaborative learning by
playing…
Imagination training, interactive games, influence on (bad) comportment ?
Intelligent car
intelligent house ? management of vital functions, recognition of visitors, spare of
water and energy, household appliances, equipment for handicaps..
Eunika Mercier-Laurent
17
Holistic Knowledge Management
An integrated system of initiatives, methods
and tools designed to create the optimal
flow of knowledge within and throughout an
extended enterprise to ensure
stakeholders success
Debra M. AMIDON ENTOVATION International
Eunika MERCIER-LAURENT
3D of KM : Technology, Economy, Social/Culture
ICAI'06
Eunika Mercier-Laurent
18
Holistic Perspective
Society
Nations/
Regions
Enterprise
Groups
Individual
ICAI'06
5
Multi-Lateral Agencies,
International Societal
Organizations
4
Countries, Consortia,
Regional Entities
3
Companies, Universities,
Government Agencies
2
Functions, Teams, Associations
Disciplines, SIG’s,
Community of Practice
1
Employees, Suppliers,
Customers, Stakeholders,
Alliance, Partners
Eunika Mercier-Laurent
19
3 levels of needs
• Individual – machine as an amplifier of human
capacity and intelligence
• Organizational
• Society
ICAI'06
Eunika Mercier-Laurent
20
Individual level
• Intelligent assistant able to
• Find relevant information and knowledge on demand
and push
• Understand documents/emails, make un abstract
• Manage my documents, files, emails using my logic
• NL dialog and capability of RT translation
• Optimize tasks (travel, shopping, event..)
• access and intelligent navigation in the world
knowledge bases (scientific, music, sport,)
• Recognize visitors, automatic vacuum cleaner,
advice…
• Tell me a joke when I am sad..
ICAI'06
Eunika Mercier-Laurent
21
Entreprise/Organizational level
The same capacity as for individual +
• Effective management of innovation process
Computers are considered as a source of K
• Sharing the learned knowledge relevant to a given
point-of-view with relevant people
• Effective management of intellectual capital, “who
knows what, who needs to know what, and how to
learn what is needed”
• Support for “business intelligence”: finding and
checking relevant information
ICAI'06
Eunika Mercier-Laurent
22
Entreprise/Organizational level
• Patent browsing for similarity determination
• Automated tools for pattern discovery and
knowledge acquisition both at the individual and
collective levels
• Tool for building collective experience of the
company
• Decision support for all professionals : diagnostic,
configuration, problem solving, process control…
• Management of global security
ICAI'06
Eunika Mercier-Laurent
23
Society
• Intelligent e-services (administration, tourism, RT
education, call centers…)
• Intelligent connections between enterprises,
university and investors 3P
• Entertainment closer to the life (intelligent
games, travel guides and tips, VR visit of
monuments with interaction (touch, smell…)
• Intelligent banking services
• Bank of knowledge and experience (health…)
• Services for older people
• Health services
ICAI'06
Eunika Mercier-Laurent
24
Challenges for AI
• Better understanding of our brain capacity
• Better use of computer capacity (K thinking
approach)
• More collective, multi-domain and multi-cultural
intelligence (1+1=11) instead of competition
• Intuitive software
• HM natural communication
• RT translation
ICAI'06
Eunika Mercier-Laurent
25
Challenges for AI
• The Convergence of bio- info- and nanotechniques (human spare parts) ?
• Extension of biological capabilities ?
• Global Innovation Support Systems
• New ways for communication
• Better conservation and use of past
knowledge and experience
• Contribution to the Sustainable Knowledge
Society
• More of fun
ICAI'06
Eunika Mercier-Laurent
26
Share our dreams, lets work our
imagination
and
Innovate our future together…
谢谢!