Transcript Application of Artificial Intelligence of for the Development Africa

Application of Artificial Intelligence for
the Development Africa
Dr Osei Adjei
Overview
•
•
•
•
•
•
•
•
•
What is Artificial Intelligence?
AI applications
Agriculture
Other significant applications.
AI paradigms
Previous applications.
Current Research and applications
Conclusions
Acknowledgements
What is Artificial Intelligence?
• Machines that mimic the behaviour of
human beings.
• Machines that can think for themselves.
• These machines can be robots, a product
inspection machine on assembly line,
machine for diagnosing diseases, software
etc.
AI Applications
•
•
•
•
•
•
Forest land mapping and classification.
Forest growth and dynamics modeling,
Spatial data analysis and modeling.
Plant disease dynamics modeling.
Climate change research.
Agriculture.
Agriculture
• An expert geographical information system
for land evaluation.
• Artificial neural networks for plant
classification with image processing.
• Simulation: Biomass growth model.
• Control of green house.
• Control of biomass.
Other significant applications
• Transportation: Planning and logistics
• Medicine: Diagnosis of cancerous cells, Xray and NMR images.
• Military: Detection of enemy aircrafts,
remote sensing.
• National and airport security: Face
recognition, fingerprint recognition, fast
DNA data analysis.
Modelling
• Complex systems can be non-linear, difficult to
formulate any mathematical expresssion and are
multi-variate.
• Traditional methods (usually statistical methods)
are not adequate to model complex system.
• AI is a useful tool for modeling such systems.
• AI produces results that even surpasses those
derived from traditional methods.
AI Paradigms
• Neural Computation (also known as
Neural networks)
• Fuzzy Logic
• Genetic algorithms
• Expert Systems
• Swarm Intelligence
Neural Computation
• Neural computation is carried out by a number
basic elements.
• The elements are modelled on real neurons in
the mammalian brain.
• The elements put together function as parallel
distributed processors.
• Thus, they have the ability to model any
complex system.
Neural network
Courtesy: http://interests.caes.uga.edu/
Fuzzy Logic
• Fuzzy Logic is a multi-valued logic that allows
intermediate values to be defined between
Aristotelian logic evaluations, where only
true/false, yes/no or 1/0 is used.
• Notions such as hot or very comfortable can be
formulated mathematically and processed by
computers.
• It provides a simple way to arrive at a definite
conclusion based upon vague, ambiguous,
imprecise, noisy, or missing input information.
Fuzzy Logic (cont.)
• Fuzzy Logic has been used in control
applications more often because the
controllers are easy to design, often
requiring few hours to develop a very
sophisticated system.
• Although the Fuzzy Logic designs are
simple their performances are usually
higher than those using traditional
methods.
Genetic Algorithms
• Genetic Algorithms are a means by which
machines can emulate the mechanisms of
natural selection.
• This involves searching high-dimensional
spaces for superior or optimal solutions.
• The algorithms are simple, robust, and general.
• GAs assume no knowledge of the search space
and can be computationally intensive.
• GAs are inherently parallel in nature and so their
execution rate increases with the number of
processors.
Expert Systems
• Expert systems represent knowledge in the form
of if-then-else rules.
• Initial stage of building an expert system is to
acquire the rules.
• This is also known as the knowledge acquisition
stage.
• Knowledge acquisition requires the interaction
between two specialists: A knowledge engineer
and a domain expert.
• The outcome of such a meeting is that a
preliminary rule base is produced.
Evaluating rules in Expert
Systems
• The rules are evaluated on trial data with errors
noted.
• Following this, subsequent meetings are held
between the two specialists to resolve all the
erroneous issues.
• The main problem associated with the
construction of an expert system is that
sometimes the domain expert may not be fully
cooperative or he/she might find it difficult to
create rules that he/she uses in his work.
• Hence, the system may not be fully realized.
Swarm Intelligence
• Swarm intelligence refers to the property of a
system in which the collective behaviours of
(unsophisticated) agents interacting locally with
their environment cause coherent functional
global patterns to emerge.
• It provides the basis which makes it possible to
explore a collective (or distributed) problem
solving without centralized control or the
provision of a global model.
• Examples of such a system are ants and bees
colonies
Applications of AI
• Security applications:
– Fingerprint applications
– Facial identification.
•
•
•
•
•
•
Inspection of manufactured parts.
Bin picking.
Visual Inspection of objects (e.g. fruits).
Medical Intelligent Tutoring Systems.
Transportation.
Medical Applications:
– Diagnosis
– NMR/X-Ray Slides Diagnosis
– Determination of mineral resources.
Current Research and
Applications
• A Data-driven Adaptive Tutoring System
for teaching Modern Standard Arabic
language
• Intelligent Multimedia Medical Assessment
Based on Hidden Markov Model.
• Detection of Drowsy Behaviour in
restrictive environment.
• Trust: Intelligent Agents for policing the
Internet.
Conclusions
• Applications range from all areas of both the industrial
and commercial infrastructure of countries.
• It has been applied in agriculture, health, medicine and
as a teaching tool in the form of intelligent tutoring
system.
• AI is not just a scientist tool but a very significant tool
that will help Africa and the rest of the developing world
to solve a vast number of their problems.
• Africa must be more involved in AI technology because it
has been successful in the developed countries.
• It is now a matured technology that can be applied
anywhere with confidence.
Acknowledgement
Many thanks to the UN, IRAC, CATS Faculty,
University of Bedfordshire for partially sponsoring my
trip to Addis Ababa to present this paper.