Tuesday, January 15, 2019

Blog 1 - How future AI will be programmed


AI or artificial intelligence is a prospect that is similar in concept to the old Frankenstein movies in that people are playing god. The completed version of an AI is one that think and act within freely being able to develop new knowledge on its own. It is essentially creating a sentient being but in the case of artificial intelligence the possibilities are limitless because they can be programmed to think faster and more logistically than humans, in addition to being given bodies making the AI more adept at certain tasks. Key things to keep in mind when programming the AI is what the desired goal is, control systems, behaviors, arbitration, decomposition, physical interfaces, and implementation (J. L. Jones 2004). Programming something this powerful is not as straightforward as telling it what is what, if that were the case no human would be able to program anything close in their or multiple lifetimes. Instead it has become more like an art being individually developed with unique tools, concepts and techniques (Michael L Brodie 1984). One possible concept for programming such a thing would be to not even have humans do it but instead have highly specific simple AI programs develop more complex ones. This could be done by building two AI, one to build and one to test. The building AI would create new AI that serve the desired purpose while the tester would be given a list of data relative to the desired AI’s purpose. Millions of AI that have been built could be sifted through using the tester using a sort of true false method. Then after the test is complete the AI with the highest success ratio could be used as a base model to create a new set of AI and this process could continue until an AI is developed with a 100% success ratio. This would allow for a much more high speed development in the future more complex AI. Other concepts include things like data structures, dynamic typing, object oriented programming and compiler self hosting  (Jones 2009). However the question arises are these really free thinking AI or just machines with a vast knowledge of the world that can not be expanded. This is something that can only be answered in the future and by one far more qualified than I am.


Works Cited
Jones, Joseph L. Robot Programming a Practical Guide To Behavior-Based Robotics. New York: McGraw-Hill Companies Inc., 2004.
Jones, Tim. Artificial Intelligence A Systems Approach. Toronto: Jones and Bartlett Publishers, 2009.
Michael L Brodie, John Mylopoulos and Joachim W. Schmidt. Topics in Information Systems On Conceptual Modeling. Cambridge: Springer-Verlag New York Inc., 1984.



Comments:

Albert, I really appreciate this post because of how relevant it is to the current times with information being valued even higher than gold. I know for certain there are times when I do a simple google search and then I start seeing ads pop up on related information. I can agree that I don't really think AI's development is something to fear but for some in order to advance a new age of technology that AI is sure to bring a leap of faith is necessary and something scary like complete exposure to an unknown source doesn't really help.

Yicheng, the topic of 3-D printing gets me excited because to me it is such a novel thing being able to design something in a modeling software and then having a machine make exactly that. It is even more impressive to see how far along the hardware is coming to the point that it can actually be used for construction purposes. I could see the final products being more lightweight but I had no idea that it would be possible to create stronger building materials with a 3-D printer. It will be interesting to see how BIM develops in order to cope with the growing demand for 3-D printed materials and how it will change manufacturing and construction.

Weiyi, the development of a more optimized HVAC system in every day housing is a concept that interests me and I look forward to its future development however I feel it is still in the early stages to truly be accessible. As an every day person I feel it is out of reach for both cost and instillation purposes so I guess I'm more curious about how this will increase in accessibility. If that could be done I think it would be incredible and the numbers you provided on how the efficiency would allow energy consumption of a 3,500 SF home to decrease to that of a 2,000 SF home emphasize why the public availability of this technology is important.


No comments:

Post a Comment

Note: Only a member of this blog may post a comment.