Pattern Recognition & Artificial Intelligence
Summary
Printed letters are visual patterns. The U.S. Postal Service uses pattern recognition machines to recognize letters and read the addresses on envelopes to sort them for delivery. Words are auditory patterns. Speech recognition software on computers recognizes words, making it possible for a personal computer to print what one speaks into a microphone, and to take verbal commands. Live video images of roads, people & cars taken by a camera mounted on a car are visual patterns. Pattern recognition software makes it possible for a car to drive itself while staying on the road and avoiding hitting people or other cars. A human face is a visual pattern. Pattern recognition software is able to scan images of people looking for known terrorists. The ability of machines to reliably recognize patterns is much inferior to human pattern recognition ability, but is improving steadily, and the number of potential applications is large. As the power of computers continues to double about every one and a half or two years, the ability of machines to recognize patterns will continue to improve.
An artificial intelligence is a machine or computer program which is able to do some of the things that currently can be done by people much better than by computers. Pattern recognition is one aspect of artificial intelligence. Other aspects include game playing, intelligent control of robots and other machines, reasoning, creativity, interacting with people by speaking and listening, problem solving, etc.
Applications of pattern recognition and artificial intelligence include the following.
- Reading Print & Handwriting - For sorting mail, reading amounts on checks, reading & processing documents, reading for the blind, etc.
- Speech Recognition - For vocal control of computers & text entry in computers, especially small portable computers with limited keyboards
- Biotechnology - Finding genes & control patterns in DNA, finding similarities in genes & in proteins, finding genetic patterns of cancers, etc.
- Robotics - Efficient effective control of robot arms, vision for robots, autonomous or semi-autonomous robots that need little supervision, bipedal humanoid robots that can walk and run safely and efficiently
- Aids for the Handicapped - Readers for the blind, guides for the blind,
- Autonomous Vehicles - For carying supplies in military convoys, for space exploration, for safer highways
- Airplane Autopilots - Automatic takeoff & landing, semi-autonomous pilotless drones
- Better Factory Automation - Machines that can operate with less human supervision, machines with vision and some intelligence that can deal with variability better than current machinery, machine inspection for product flaws, etc.
- Anti-Terrorism - Recognizing terrorists, recognizing bombs, guns, etc.
- Internet Search Engines - Search engine companies are trying to create search engines that understand to some extent the documents they are cataloging, and respond to natural language queries instead of just looking for keywords.
- Internet Routers - Internet routers use artificial intelligence for routing messages
- Understanding How the Human Brain Works - Studying ways to produce machine intelligence often sheds light on how the human brain works. Many advances in psychology, neurobiology, and artificial intelligence have resulted from applying knowledge from one of these fields to another.
Research in Pattern Recognition & Artificial Intelligence at the University of Arkansas
The following Electrical Engineering faculty are interested in pattern recognition or artificial intelligence. Details on their research can be seen by clicking on their names.
- Randy Brown has done research for the U.S. Postal Service on machine recognition of addresses for mail sorting. His current interests include machine vision, reinforcement learning, artificial neural networks, and application of insights from neurobiology to pattern recognition and artificial intelligence.
- Magda El-Shenawee has applied simulated evolution methods to the pattern recognition problem of microwave imaging breast tumors.
- Vasu Varadan is interested in pattern recognition problems associated with microwave imaging. She is researching both security applications and medical imaging.
Courses for the Pattern Recognition & Artificial Intelligence Specialty Areas
Recommended Undergraduate Elective Courses
MATH 3083 Linear Algebra - Take as a Math/Science Elective
CSCE 2143 Data Structures
CSCE 4613 Artificial Intelligence
ELEG 5653 Artificial Neural Networks
ELEG 4403 Control Systems
Additional Graduate & Undergraduate Courses
CSCE 3313 Algorithms
ELEG 5443 Nonlinear Systems Analysis and Control
ELEG 5453 Adaptive Filtering and Control
ELEG 5473 Intelligent Transportation Systems
