Recent Trends in Artificial Intelligence

spineunkemptAI and Robotics

Jul 17, 2012 (5 years and 3 months ago)

523 views

1  Journal of Organizational Leadership & Business                                                     Summer 2007 
 
 
Title: Recent Trends in Artificial Intelligence 
Author: Breanna Merenda 
Professor: Charles L. McDonald, Jr., Ph.D. 
Class: MIS528 Emerging Technologies in MIS 
Semester: Spring 2007 
 
Assignment: 
 
Students in TAMU-T’s graduate level Emerging Technologies in MIS class of spring 2007 were
assigned to conduct research on an emerging technology topic of their choice, submit a
proposal with references, and, if approved, develop papers that satisfied specific criteria. Each
paper should contain an introduction, a summary, and a section addressing the significance of
the topic. The Turnitin site was used to evaluate the paper for plagiarism and provide an
originality report to the student. To receive credit, papers had to satisfy academic integrity
issues, be grammatically correct, free of spelling errors, meaningfully organized, and content-
rich.  
Abstract: 
 
Artificial Intelligence (AI) is the use of science and engineering to grant computers the ability to
perform tasks by mimicking human intelligence. The applications of AI technology range from
simple tasks, such as typo corrections, to advanced tasks, such helping dementia sufferers.
There is no unifying theory in AI; rather, the field is governed by a myriad of theories about how
to represent human thought in a machine effectively. The field of AI is diverse and
encompasses many applications. Health care industries are incorporating AI into their
databases and are researching how AI can benefit dementia sufferers. Educational institutions
are utilizing AI in classrooms and online learning and the United States military is applying AI for
training purposes. AI is utilized by day traders on Wall Street and by politicians; further more,
businesses are gaining a competitive advantage by protecting workers and reducing errors
because of AI applications.  
 
2  Journal of Organizational Leadership & Business                                                     Summer 2007 
 
Recent Trends in Artificial Intelligence
 
Introduction


Artificial Intelligence (AI) is defined as the use of science and engineering to give
computers the ability to perform tasks by mimicking human intelligence. The lack of a
unifying theory for AI has led to the field being governed by a multitude of theories about
how to represent human thought in a machine effectively. The field of AI is diverse and
involves applications that are used by the military, health care professionals, businesses,
and consumers.
Summary
AI applications attempt to make computers accomplish complex tasks that normally
require human intelligence. Some tasks are difficult for humans to do but are easy for a
computer to accomplish, such as advanced mathematics. Some tasks are easy for humans to
accomplish, but can be very difficult to replicate in machines, such as facial recognition.
Scientists differ on their perceptions of the similarities and convergence between human and
artificial intelligence. Some scientists believe the two systems use different methods to make
decisions; for example, computers perform millions of computations to make a decision while
humans use pattern recognition or intuition. Other scientists believe human intelligence and AI
are the same because they use the same processes to recognize patterns. These scientists
believe the only differences lie in the speed at which the two systems make decisions and how
much a person understands and assesses a situation. There are various reasons why people
study the automation of human intelligence. Some researchers develop AI applications to
research human intelligence; for example, research is being conducted to refine psychological
theories by using programs that replicate parts of human behavior. Other researchers develop
AI applications to create smarter machines by utilizing human reasoning to create novel
techniques for solving complicated problems.
AI systems have several essential characteristics, all of which rely on observation, logic,
and pattern recognition. The systems utilize built-in interference machines, which adhere to
established logical patterns. When a system is introduced to a problem or situation, it
recognizes the problem derived from patterns or parameters a developer previously established.
The system reacts to these patterns based on actions that were programmed by the developer
of the system. The system accomplishes this with if-then reasoning; for example, computers
compare specific situations with certain results. Actions made from this reasoning are
3  Journal of Organizational Leadership & Business                                                     Summer 2007 
 
categorized as classifiers, which classify an object, or controllers, which act as a response to
the situation.
The field of AI is governed by a multitude of branches and concepts, such as logical AI,
search, pattern recognition, representation, inference, planning, epistemology, heuristics,
learning, and genetic programming. Logical AI is the most basic principal in the study of AI.
Programs use mathematical logic to represent its knowledge of the environment, its goals, and
its current situation. The program decides how to act by inferring what procedures are suitable
for achieving its goals. Search is an important aspect of AI because it inspects all of the possible
courses of action by searching through a theorem-proving program. Pattern recognition is used
by programmers to allow AI systems to compare what it is observing with a pre-programmed
pattern. Representation involves how details about the environment can be represented to a
computer, which is usually with languages of mathematical logic. Non-monotonic inference is
the “default reasoning in which a conclusion is to be inferred by default, but the conclusion can
be withdrawn if there is evidence to the contrary.” Programs use non-monotonic inference to
make decisions when mathematical logical deduction is inadequate. An example of how
humans use non-monotonic inference occurs when a person sees a bird and infers it is capable
of flight. If he discovers it is actually a penguin, he can reverse his original inference. AI
programs use connectionism and neural nets to learn from experience; unfortunately, learning
systems are currently very limited in their ability to represent information. The Kelley
Engineering Center at Oregon State University (OSU) is working on a computer that plays
online video games and uses information learned from those games to play other games.
Epistemology studies what types of information are necessary for solving problems; conversely,
ontology studies the basic properties of objects interacting with programs. There are two
aspects to heuristics, which is “a way of trying to discover something or an idea imbedded in a
program”: heuristic predicates and heuristic functions. Heuristic predicates are used by
programs to decide which of the options presented is better. Heuristic functions are used by
programs to decide how far each option presented is from its goal. The newest advancement in
AI technologies is in genetic programming, which incorporates neural networks and genetic
algorithms to endow computers with “nonlinear decision making processes.” Genetic
programming was a reaction to the previous belief that programmers were required to teach a
computer a list of sequential rules so that it can perform a task. Neural networks allow
computers to generate their own rules and assumptions by testing thousands of random
4  Journal of Organizational Leadership & Business                                                     Summer 2007 
 
processes and sequences. Genetic algorithms allow programs to evolve by allowing the rules
from neural networks to compete and combine to make successful outcomes.
Significance of Topic
The most recent innovations in AI technology are in health science fields. Researchers
at the Ben-Gurion University created a program, called KNAVE-II, which provides medical
information to doctors. KNAVE-II incorporates “knowledge-based interpretation, summarization,
visualization, interactive exploration of large numbers of distributed time-oriented clinical data,
and dynamic sensitivity analysis.” Doctors who are performing follow-ups on bone marrow
transplant patients are using KNAVE-II to answer specific questions. Doctors may ask the
program to provide the names of male patients who have had difficulties since last year, instead
of referring to paper charts and Excel worksheets. Scientists are researching how AI can help
dementia and Alzheimer’s sufferers. People who suffer from advanced dementia are often taken
from their homes and placed in nursing homes. Scientists are creating AI-driven automated
systems to allow dementia sufferers to stay in their homes and ease the burden placed on
caregivers. Smart bathrooms assist individuals by using verbal and visual cues. If an individual
is confused about how to perform a task, then a video, which is located above the sink, guides
the person through the correct steps. The video provides the exact layout of the individual’s
bathroom and shows hands performing the specific steps necessary to perform a specific task,
such as hand washing. There is a camera mounted on the ceiling, which follows the movements
of the person as he completes the task. The program offers verbal cues throughout the process
if the person becomes confused. The program learns from experience; for example, some
people will not require many verbal and visual cues but others may need to hear their name
every few minutes. The program will then tailor the frequency of cues based on actions from the
user; moreover, since everything is automated, the program maintains the patient’s privacy.
This program not only guides dementia sufferers through basic hygiene tasks but also reminds
people to take their medicine and shows them the necessary steps. The results from trials of
this technology show the impact it has on making the lives of dementia sufferers easier. There
was a 25% increase in hand washings that were performed without the aid of a caregiver. The
time it took patients to wash their hands decreased from 10 minutes without the program to 30
seconds with AI.
The U.S. military is utilizing AI to help train its soldiers and maintain its position as the
preeminent military force in the world. Mission Rehearsal Exercise (MRE) is a complex
simulator created for the military to aid in the training of soldiers who are involved in combat,
5  Journal of Organizational Leadership & Business                                                     Summer 2007 
 
humanitarian, or peacekeeping missions. The Institute of Creative Technologies (ICT) acted as
a contractor for the military to create MRE. ICT collaborated with authors and playwrights from
Hollywood, Army strategists, and professors from the University of Southern California (USC).
Each soldier is presented with a 5-minute scenario projected on a 150-degree screen. The
audio is 10.2-channel, which creates sound effects that will shake the floor and walls if a sound
is loud enough. The soldier’s sense of smell is incorporated into the simulation; for example, a
moderator can infuse a room with the smell of burning charcoal. Soldiers can touch objects in
the simulator and solicit responses from objects or people they encounter. Voice-recognition
technology is incorporated into the simulator so that soldiers can converse in any language with
the simulated people. Researchers at ICT programmed the AI to respond in unpredictable ways
to ensure the experience is realistic. The unpredictability of the program also ensures the
soldier does not face the same experience every time he tries the simulator. The simulator can
be changed according to the soldier’s mission; for example, soldiers in Afghanistan switched
from combat training to humanitarian training, such as disaster aid. The Office of the Joint
Chiefs of Staff recently directed the creation of strategic computer and video games to aid in the
team-fighting approach training method for soldiers. These strategic games allow combat teams
to rehearse missions prior to arriving at their mission location. The Defense Advanced Research
Projects Agency (DARPA), under the direction of the U.S. Department of Defense, is funding
research at OSU. DARPA spends approximately $2 billion each year “to fund high-risk but
potentially high-yield experiments around the country,” especially those that apply AI to
surveillance.
Educational institutions are incorporating AI into their lesson plans to teach skills that
range from proper typing techniques to problem solving. The Artificial Intelligence Mark 2 (Ai2)
robot is Microbric’s newest educational tool. The Ai2 robot can communicate with and learn from
other Ai2 robots, be directed via a remote control, negotiate obstacles, and interact with
student’s input on an activity mat. AI software is being incorporated into eLearning. Two second-
generation eLearning programs, uSim and uLearn, were deployed recently by software
engineering company uMind. The two programs can “estimate, control, and anticipate learner
behavior” and help students learn via a series of real-time modifications and adapted feedback.
These programs are innovative in the field of eLearning, which previously only provided
information to students, instead of engaging them actively. Preliminary results from the
implementation of uMind programs show a 50% decrease in learning time and a 35% increase
in information retention.
6  Journal of Organizational Leadership & Business                                                     Summer 2007 
 
Wall Street, politics, businesses, and consumers are benefiting from AI innovations. The
Apama Algorithmic Trading Platform is a program used by day traders to pick stocks. The
program uses trading algorithms and neural networks to search historical trends and current
stock prices to provide users with rules that will help them choose the most beneficial stock.
Previous stock analysis programs had pre-programmed assumptions and rules; however,
programs with neural networks can create new assumptions, learn from previous mistakes, and
predict future events. Researchers are studying how to quantify items, such as newspaper
headlines, to be utilized to predict changes in the market. The Massachusetts Institute of
Technology (MIT) is using this technology to study how positive and negative words in an article
affect a company’s stock. One-third of all stock trades in America are powered by automatic
algorithm programs. RegentAtlantic Capital created iRebal, which evaluates the investments in
client’s portfolios to decide whether buying or selling is the best decision. Some AI tools help
politicians locate supporters; for example, data mining technology searches electronic
information and the Internet to uncover trends in people’s preferences. Microtargeting software
allows politicians to analyze 50 to 60 points of data to obtain information about a community’s
voting trends or behavior to find the exact location of likely supporters. Businesses use AI for
predicting machine failures and detecting employee fatigue. Cooperative Research Centre
Mining (CRCMining) created a program based on neural network concepts that can predict
when a machine will breakdown, which allows businesses to schedule preventative
maintenance. The program also offers diagnoses for problems. CRCMining also created a
technique to detect how fatigued machine operators are by using a “self-training artificial neural
network” that deciphers their level of drowsiness by monitoring their brainwaves.
The applications of AI technology range from simple tasks, such as typo corrections, to
advanced tasks, such as making dementia sufferers lives more manageable. While there is no
unifying theory that governs AI, each branch has succeeded in creating smarter machines and
smarter people. Tasks that were once difficult and time consuming are now completed in a
matter of seconds because of the speed and efficiency of AI-driven technologies. The military
can train soldiers for combat without risking their lives and educational institutions can maintain
their student’s attention by having robotic teaching aids. AI concepts are updated and innovated
frequently, which enables researchers and scientists to test their own limits with those of
computers.

7  Journal of Organizational Leadership & Business                                                     Summer 2007 
 
References:

“A.I. in a brave new world.” (2007, February). Australian Mining. Retrieved 7 April, 2007,
from the LexisNexis Academic Search Premier database.

Benchetrit, O. (2007, March). “uMind artificial intelligence software poised to take over
eLearning world.” Canada NewsWire. Retrieved 7 April, 2007, from the
LexisNexis Academic Search Premier database.

Berg, J. (2006, October). “Meet your new officemate.” Campaigns & Elections, 33.
Retrieved 7 April, 2007, from the LexisNexis Academic Search Premier
database.
Duhigg, C. (2006, November). “A smarter computer to pick stock.” The New York
Times, 1. Retrieved 7 April, 2007, from the LexisNexis Academic Search Premier
database.
Graven, M.P. (2001, September). “Artificial intelligence: The computer genius.” PCMag.
Retrieved 9 April, 2007, from
http://www.pcmag.com/print_article2/0,1217,a=10071,00.asp

Lair, P. (2006, November). “Military things big with help from OSU, other universities.”
Associated Press Financial Wire. Retrieved 7 April, 2007, from the LexisNexis Academic
Search Premier database.
“Life sciences; Research conducted at Ben-Gurion University, Faculty of Engineering
Sciences has updated our knowledge about life sciences.” (2007, January).
Physician Law Weekly, 203. Retrieved 7 April, 2007, from the LexisNexis
Academic Search Premier database.

McCarthy, J. (2004, November). “What is artificial intelligence?” Stanford Computer
Science Department. Retrieved 29 March, 2007, from http://www-
formal.stanford.edu/jmc/whatisai/

“New aspirating smoke detection technology in USA is in a ‘class’ by itself’ ClassiFire®
perceptive artificial intelligence technology continuously monitors for changes.”
(2007, March). Retrieved 3 April, 2007, from the LexisNexis Academic Search
Premier database.

Novak, L. (2006, October). “Build your own next generation in AI; A new learning path.”
The Advertiser, 14. Retrieved 7 April, 2007, from the LexisNexis Academic
Search Premier database.

Sieberg, D. (2006, December). “War games: Military training goes high-tech.” CNN Sci-
Tech. Retrieved 12 April, 2007, from
http://archives.cnn.com/2001/TECH/ptech/11/22/war.games/index.html

8  Journal of Organizational Leadership & Business                                                     Summer 2007 
 
Ubelacker, S. (2007, February). “Artificial intelligence to help dementia sufferers; ‘Smart
bathroom’ allows those affected to live at home longer, gives caregivers a
break.” The Globe and Mail, A19. Retrieved 7 April, 2007, from the LexisNexis
Academic Search Premier database.

“What is Artificial Intelligence?” (2007). Tech-FAQ. Retrieved 3 April, 2007, from
http://www.tech-faq.com/artificial-intelligence.shtml