CIS 140 Lecture Notes - Lecture 24: Artificial General Intelligence, Speech Recognition, Exponential Growth
20 Dec 2018
School
Department
Course
Professor
Document Summary
Currently ai really good at really specific tasks. Meanwhile calculus is hard to us, but easier to do on machines than image recognition. Ai- moving from a complex equation -> anns which is more a probabilistic model. Speech recognition- ann to decipher speech but the appropriate responses are still hard coded into the program. Moores law #1 (original)- how many transistors you can put on a chip- pretty much has ended. Moore"s law #2 -increased computer power per dollar: doesn"t focus on the decreasing size of things, bc tech improvements involves switching the components. Weak agi: a lot of programs now, ex: guided surgery. When computers are better at everything than people. What could a smart computer do: make sure computers do what you want, not what you told it to do, ex: get rid of wars- computer"s solution may just to be get rid of everyone. Pros: will be able to use computers to do things, helps us.
