Majored in computer science (focused on artificial intelligence), minored in art (focused on graphic design and 3D modeling).[spoiler]Now I teach English as a second language.[/spoiler]
English
-
That twist tho
-
That's life. You never know what direction it will take you.
-
Yeah I feel ya
-
Tell me about artificial intelligence computer science, I'm pretty interested in doing something in that field.
-
Did some work with simple robotics, neural nets and parallel processing. The robotics was mostly straightforward but neural nets are quite a bit different from traditional programming. In traditional programming you create an algorithm that takes in input, processes it and gives you output. Parallel processing is having several algorithms running at once, each doing a different task so that the work can be divided up and completed more efficiently. The way neural nets work is more in line with a biological brain. You have several nodes (neurons) that take in information, process it and pass it on (or not pass it on). At the end of the line you have your output. By giving multiple inputs and expected outputs, the neural net can make guesses as to how information should be processed then check to see if it was right. The machine can then find which patterns for processing information were the most consistently successful. I'm obviously simplifying things but that's the gist of it. A good example of a real world application would be in language translation. Using traditional programming you'd have to program in every word, every grammatical structure and how they coincide. This would be excruciatingly complex, and sometimes you could run into problems where languages have no equivalent (for example, in Chinese there is no simple "yes" or "no", but rather they answer with an affirmative or negative "Do you want to eat" --> "I want to eat./I don't want to eat."). A neural net, on the other hand, would take your input, consider key words and phrases, look for similar usages of those key words and phrases in the second language and then give you a predicted output. For instance, if you said "I am riding a horse," it would look at sentences and try to find previous successful translations the given words, then try to look for examples of how they are associated in the second language and come up with its best estimate of an accurate translation. The more translations it makes (as long as it knows it succeeded failed or got close), the better it gets at refining it's ability to find accurate output.
-
Very interesting. Thanks.
-
No problem. Hope you enjoy it.