Budny 10: 00 L19 Disclaimer—



Download 20.88 Kb.
Date28.01.2017
Size20.88 Kb.
#8980

Budny 10:00

L19

Disclaimer—This paper partially fulfills a writing requirement for first year (freshman) engineering students at the University of Pittsburgh Swanson School of Engineering. This paper is a student, not a professional, paper. This paper is based on publicly available information and may not provide complete analyses of all relevant data. If this paper is used for any purpose other than these authors’ partial fulfillment of a writing requirement for first year (freshman) engineering students at the University of Pittsburgh Swanson School of Engineering, the user does so at his or her own risk.
ARTIFICIAL INTELLIGENCE: GENETIC PROGRAMMING
Tyler Vogel (tjv21@pitt.edu)



ARTIFICIAL INTELLIGENCE AND REVERSE ENGINEERING THE BRAIN
Description
Artificial intelligence (AI) has been a concept that engineers have been working towards since the dawn of computer engineering. AI is a computer’s ability to imitate human intelligence, such as thought, behavior and the ability to learn [1]. It is one of the greatest challenges of computer science and engineering to invent a machine that can simulate the cognitive functions that a human possesses, or even simulate a greater intelligence then that of a human.

The earliest programs that tried to venture into AI were small projects, such as checker games, math word problem solvers, and basic speaking programs. Computers were slowly getting closer and closer to having the same intelligence capacity as a human in the 90s, but when suddenly science hit a wall and progress slowed down [2]. Computer scientists were not able to recreate intelligence that was as complex as the human brain, and they are still not able to this day. This is the next big leap in AI: reverse engineering the brain.

Reverse engineering the brain is just the scientific process that allows engineers and scientists to discover the secrets of the complex human brain [1]. The knowledge acquired by this difficult project can be used to help create a sophisticated artificial intelligence, able to learn, grow, and act like a human. There are many obstacles in the way of this achievement, including complex coding, not fully understanding how neurons transmit information, and physical circuit complications.
Significance
Artificial intelligence has an infinite amount of uses. It is used in manufacturing, personal computers, video games, robotics, and even in some medical practices [1]. Having the ability to reverse engineer the brain and bring AI into a whole new era would do more than just create fancy robots (although that is a nice perk). One of the major opportunities created by reverse engineering the brain is better biotechnology, which is technology that can improve normal human functions. This includes helping the disabled with their disabilities, mental and physical, and also expanding the limits of any human being, humans can become smarter, faster, better, and stronger.

AI doesn’t only hold the key to creating a superhuman, it also can improve the core processing of our computers. Intelligent programs can lead to more accurate medical diagnosis, speech recognition, autonomous vehicles, artificial assistants, image recognition, mathematical theorems, etc. [1]. The list can go on forever. Right now, the world is at the beginning of the digital age, and improved supercomputers can only help society and engineering to advance and grow.

The Accreditation Board for Engineering and Technology defines engineering as “the profession in which a knowledge of the mathematical and natural sciences gained by study, experience, and practice is applied with judgment to develop ways to utilize, economically, the materials and forces for the benefit of mankind [3].” Artificial intelligence is somewhat of a debated topic concerning whether it would benefit mankind or destroy it. People are scared that if we create intelligent machines, our jobs and lives will be taken away. CGP Grey’s educational video, “Human’s Need Not Apply,” describes how machines will take away both blue and white collar jobs from humans. Intelligent machines would not make as many mistakes as humans, are not as costly as human workers, and would eventually become easily replaceable [4]. At the same time AI makes life easier for some, it also makes it harder for others.

It has been a dream of mine to work on and perfect artificial intelligence. I know that eventually a machine could do the job I want to do after graduating, but I want to see continued progress in the computer science field. Funny enough, a movie called “Chappie” is one of the biggest inspirations for me to support the development of AI. It is about a particular police droid that is stolen and given new programming that gives it the ability to feel and think for himself. If I ever get a chance to work on AI, a Chappie-like robot is going to be one of the first things I push for. Another inspiration is CGP Grey, an Oxford Professor who also makes educational videos on the side. He promotes the growth of artificial intelligence, and a few of his videos explain why AI is a positive rather than a negative development.


AUTOMATIC PROGRAMMING
Description
Artificial intelligence is not just one big concept, but instead is a field of many parts. One large part of AI is automatic programming, which is when computer programming is done by a machine. Automatic programming isn’t a stagnate goal, but a moving one that gets bigger and more difficult to overcome as expectations get higher [5]. Much of what was originally conceived of as automatic programming was achieved long ago, such as a compiler program or an assembler used in a modern programming language.

The expectations of an “automatic programmer” today are somewhat unrealistic, such as having a machine create a completed code for any task that the user inputs into it. Researchers at MIT and other schools are still trying to accomplish this task despite how ridiculously hard it seems. Instead of taking up a task this large, many languages such as Java and C++ have instead made programming slightly automatic with tools such as natural language, shortcuts, and even GUI’s (graphical user interfaces) [5].

For a more complete automatic program, computer scientists need to make a system that will have three key features. First, the program must be end-user oriented, communicating directly with end users. It also will have to be general purpose, working as well in one domain (computers, softwares, etc.) as in another. Finally, the system has to be fully automatic, requiring no human assistance.
Significance
Automatic programming is the gateway to a futuristic, sophisticated artificial intelligence, but is also helpful in our daily lives now. Many people do not think about how our computers still run off of 0’s and 1’s, due to the fact that all coding uses short cuts, so programmers can use words instead of writing out every single digit. Our computers and smartphones wouldn’t be anywhere near as efficient as they are now if automatic programming did not exist [6]. Today, society runs on the internet and programs to function, so improving automatic programming would be a plus for everyone.

Engineering relies heavily on automatic programming as well. Just think about how many computer programs are used on just one simple project (SolidWorks, MatLab, C++, Python, etc.) and how hard it would be to accomplish anything in the engineering world without these. Automatic programming makes all of this possible and efficient, which is why it is so essential to the advancement of engineering [6].

Automatic programming has also been significant to my educational career as well. In high school, I had multiple engineering and computer courses. The shortcuts I learned while programming and designing helped me accomplish my projects in these classes. If anything, I want to improve automatic programming so my coding can be less time consuming and more user friendly.
GENETIC PROGRAMMING INC.
Description
Genetic Programming Incorporated is a company that specializes in automatic programming. Their goal is to create a programming method, called genetic programming (GP), which is “an automated method for creating a working computer program from a high-level problem statement of a problem [7].” This means that GP starts from a high-level statement of “what needs to be done” and automatically creates a computer program to solve the problem. This is done by having a program that grows and learns.

GP starts with thousands of randomly created computer programs. The group of programs slowly evolve over generations following Darwin’s theory of evolution. GP uses a fitness measure to test how well each program does its task, and decides if the program needs to be improved, thrown out, or is complete [7]. Sometimes, a single program must be tested multiple times with different initial conditions or different environments. Overall, genetic programming is a learning program that improves over time all by itself, also known as an “intelligence explosion [8].” It’s a perfect example of an automatic program, and also a core component to artificial intelligence.


Significance
A learning program like genetic programming has numerous applications in artificial intelligence, but Genetic Programming Inc. wants to focus on black art problems and programming the unprogrammable. Black art problems consist of the difficulties of automated synthesis of analog electrical circuits, controllers, antennas, networks of chemical reactions, optical systems, and other areas of design [7]. Anything moving towards automation is great for making life easier for engineering and society. Programming the unprogrammable according to Genetic Programming Inc. is “the automated creation of computer programs for unconventional computing devices such as cellular automata, multi-agent systems, parallel programming systems, etc. [7].” This is very useful for nano-bots and swarms of drones or robots, since they will be able to learn from their surrounding bots and follow their actions, similar to how flocks of birds or schools of fish work in the real world. Genetic programming is the start of artificial intelligence and could be one of the most important technological steps in the digital age.
THE ETHICS OF THE SINGULARITY
There is a concern with artificial intelligence that is important to mention: the singularity. The singularity is a point in the future when AI reaches and then surpasses human intelligence. Some people see these super intelligent machines working for humans and benefitting for humanity, while others worry that these machines threaten humanity [8]. They are scared that intelligent machines might begin replacing human workers and end up replacing human beings altogether. I believe, however, intelligent machines will only benefit humanity, but the fear of intelligent machines might slow down progression of AI.

SOURCES
[1] “Reverse-Engineer the Brain.” NAE Grand Challenges for Engineering. 2016 Accesses 10.30.16. http://www.engineeringchallenges.org/challenges/9109.aspx [2] “AI Overview” AITopics. Accessed 10.30.16. http://aitopics.org/topic/ai-overview

[3] “Accreditation.” ABET. Accessed 10.30.16. http://www.abet.org/accreditation/

[4] CGP Grey. “Humans Need Not Apply.” CGP Grey YouTube Channel. 08.13.14 Accessed 10.30.16 https://www.youtube.com/watch?v=7Pq-S557XQU

[5] Charles Rich, Richard C. Waters “Approaches to Automatic Programming.” Mitsubishi Electric Research Library. 2013. Accessed 10.30.16. http://www.merl.com/publications/docs/TR92-04.pdf

[6] “Automatic Programming” AITopics. Accessed 10.30.16. http://aitopics.org/topic/automatic-programming

[7] “What is Genetic Programming (GP)?” Genetic Programming Inc. 2012 Accessed 10.30.16. http://www.genetic-programming.com/#_What_is_Genetic



[8] Tsin Yen Koh. “Intelligent Machines vs. Human Intelligence: Overview.” University of Pittsburgh Library. 2015 Accessed 10.30.16. http://web.a.ebscohost.com/pov/detail/detail?sid=a1b4572d-89ac-454f-ad99-122a5c99b612%40sessionmgr4009&vid=0&hid=4104&bdata=JnNpdGU9cG92LWxpdmU%3d#AN=112140580&db=pwh
ACKNOWLEDGMENTS
I want to thank Angie Farkas for helping me edit my paper. Without her opinions, my paper would have been choppy and hard to read.

University of Pittsburgh, Swanson School of Engineering

11.01.2016

Download 20.88 Kb.

Share with your friends:




The database is protected by copyright ©ininet.org 2024
send message

    Main page