Clearing the Water Around A.
I.
Big Data or Big Hype?Kal LemmaBlockedUnblockFollowFollowingMay 2Nearly everyone today has been experiencing some effects and new ideas about artificial intelligence.
Most companies, banks, retail stores, etc, are focusing on ways aritificial inteligence will expand their market and lead them to more successful ventures.
I’m sure you’ve all once had or heard a conversation with people with no technical knowledge about artificial intelligence apprehensivley talking about how their lives are going to be changed by it.
I’m sitting in this hole in the wall breakfast place by Wall Street while I’m writing this and experiencing it right now.
Just picture a couple of classic NYC executives talking about the A.
I.
strategies they’ve started implicating at their firm, “It’s going far better than we assumed, I’ve got no-clue what’s going on, I just hope we’ll be able to start making staff cuts ASAP.
”Many start ups today are trying to slap A.
I.
right on their face, trying, and usually succeeding, to generate large amounts of funding.
There is this large investment into a field that people see as todays ‘gold rush’, but there is also a large murkey public understanding of how it works.
ALGORITHMS ARE THE LEGOS“An algorithim is a set of step-by-step instructions so explicit even something as literal-minded as a computer can follow them.
”By itself, other than impressing or intimidating people just from the sound of it, an algorithim alone is basically no smarter than a power saw (A.
I.
solving power saw kickbacks Dustin).
Try and think of of something that just does one thing extremely well, repeatedly.
Its very useful for like sorting lists of numbers, or searching the internet for pictures of mother of dragons.
But once we start chaining lots of algorithims together in clever ways, we can start to play with artificial intellegince.
An A.
I.
thats just a domain-specific illusion of intellegent behavior.
Think about human evolution in how we really started to evolve once our primitive ancestors started to work in collective groups to accomplish their goals, (think of nature as being a potential parameter with the goal of survival).
WHATS HAPPENING BEHIND THE SCREENLet's use one A.
I.
example that is currently being used millions of times daily to give people access to whatever information they are seeking.
“Ok Google, Where can I get the best Thai food nearby?”A specific algorithim will take my raw recorded sound waves and transform it into a digital signal.
Another Algorithm will try to translate that digital signal of my question, into a string of English phonemes: perceptually distinct sounds.
(besté-[taį]- fuid) The process of stringing phonemes is reasonable with single words but extremely difficult with sentences.
This Algorithim uses linguistic approaches to look for ‘distinct unit sounds’, using some interesting linguistic theories.
The next algorithm will segment these phonemes, “besté-[taį]- fuid”, into the words it thinks I said.
It would most likely be just part of my sentence that it can best interpret, “best thai food”.
The result would then be sent to a Search Engine.
The Search Engine is a crucial part in where the research for the interpreted string happens.
It itself is just a huge pipeline of a bunch of other algorithms that processes the query and sends out an answer.
This answer to my question will, you guessed it, go yet to another algorithim that will format my response into a coherent English sentence.
Finally, the last algorithm will verbalize that sentence in a non-robotic, human-like way; “I found a few Asian places near you.
”Thats pretty much the concept of how nearly every AI system works: self-driving cars navigating the streets, roombas that scan your living rooms, social media feed and suggestions, online advertisment networks and many more.
SUM UPA simplified understanding of what Aritficial Intelligence does and what I have explained follows that same “pipeline-of-algorithims” above.
A.
Takes in Data from specific domains.
B.
Performs a chain of calculations.
C.
Gives out our prediction or decision.
TWO DISTINCTIONS OF ALGOS USED IN A.
I.
These algorithims, rather than working with certainties, they typically deal with probabilities (like saying “I’m 91% sure that x happened).
Unlike traditional algorithms (where a programmer fixes those instructions ahead of time in order for it to run and know what it’s doing) , A.
I.
algorithms instructions are learned by the algorithim itself.
It does so directly from the “training data” (that allows it to see thousands to millions of examples of each category) and learns to distinguish inputs from outputs.
CONCLUSIONA.
I.
is rapidly improving and the applications for it in the future are practically invevitable.
We may or may not open pandora’s box as we continue on with more and more data becoming available, our hardware technology continually advancing, and whether or not we have ethical-minded people helping design it.
For artificial intelligence, the role of the programmer isn’t to ‘tell’ the algorithim what to do, but rather find ways to tell it ‘how’ to train itself what to do: through the data and rules of probability.
Works CitedPolson, Nicholas G.
, and James Scott.
“Intro.
” AIQ: How people and machines are smarter together.
New York: St.
Martin’s P, 2018.
.