7/19/2015

Artificial Intelligence - Are We There Yet?

Are we there yet?  No, not yet.

Are we there yet now?  No, we're not there now.

Are we there yet?  No!

Artificial Intelligence has been a thorn in humanities side since the mid 1900's.  The people who developed the science of AI lacked the processing power to fulfill their goals.

A half century later, we've got a lot more processing power, a lot more public awareness and a lot more funding.

So a lot of smart people are investigating the subject of Artificial Intelligence.

Here's a good video discussion of the current state of AI:

Artificial Intelligence: Are we engineering our own obsolescence? https://youtu.be/u0R6pV_aeLc

Universal Encyclopedia
The "History of everything".  Who did what when.  Facts.  Events.  People.  Places. Time.  An example, search engine companies like Google, Bing, Yahoo, etc.  They are well on their way to storing the history of everything.  And their databases continue to grow daily.  Search capabilities exist on every device connected to the internet.

Predictive Analytics
Machine Learning is capable of predicting future events based on past criteria fed into Neural Networks.  These devices work by training Models in specific domain knowledge, refined over time for better accuracy and are able to sift through tremendous amounts of data with a good degree of accuracy. A good example of predictive analytics is Credit card applications.  The computer bases the credit worthiness of a specific individual by comparing his or her demographics against a trained model of prior applicants over time.  This technology has been around for decades.

Personal Assistants
Programs on the computer or smart phone that learn the behavior of the individual using the application.  Preferences, places of travel, history of searches and browser trails, the application keeps track of schedules, able to inform its users to alter a course of action based on given parameters in real time, these assistants are only going to get more granular, accurate, mainstream and built into other devices.

Deep Learning
How do you become proficient at driving a car?  You learned.  And how was that accomplished?  You may have read a book, took a class, watched others and eventually got behind the wheel and pressed the gas peddle, and perhaps the brakes.  And at some point, you ventured on to the road and interacted with other cars.  And every time you get behind the wheel of a car, you reference all the knowledge you've accumulated over time, and put that into action and make adjustments as necessary.  A reckless car cuts you off in traffic, you store that in your memory banks for next time, to ensure your safety going forward, essentially tweaking your model of driving a car.  So Deep Learning machines, using tremendous processing power and volumes of data, are also learning specific domain techniques.  They are "Learning".  A machine that can learn does so by unsupervised techniques.  It runs thousands of simulations, learns and tunes itself over time.  Eventually these machines will be able to cross domains by leveraging it's current skill set.  Perhaps an interconnected network of various domain  knowledge machines, constantly learning through new data fed in by input devices, tweaking the models and dispersing that information back out to external locations, machines or devices.

Conscious Computer Beings
When computers become aware of themselves, they will have crossed the line from bits and bytes and electronic switches into a living being.  This event is commonly referred to as the "Singularity".  Creating a living being from parts seems unlikely and far off into the future.  It may or may not be possible.  Perhaps more likely would be the combination of a living being integrated with some type of machine to augment the physical form.  This would accelerate the strength of the person into a hybrid form and have obvious advantages over mere humans.  Perhaps faster processing power, longer lasting memory, better mathematical capabilities, better eyesight or hearing or smelling.  Perhaps extended life with humans living additional 50 to 100+ years longer than traditional ages.

Robotics
There are some Robots that exist today that are very capable of various  tasks.  Some are agile and can transport themselves over difficult terrain.  Some don't have the traditional human form and perform various task in assembly lines.  Some have very distinguished features and can reflect facial expression that appear to be based on emotions.  There are drones that can fly great heights and distances with great speed and agility.

So, do we have Artificial Intelligence?

Well, it depends on how we define it.  Some would say that we already have a basic form of select sub-fields of AI.  Others would say that no computer has passed the "Turing Test", in that the judges can not differential between an AI  and a human.  Some people think that reaching the Singularity is the only true form of AI.

What should the definition be?  There is such pressure to produce artificial intelligence.  The clock is ticking.  Are we there yet?

We're definitely getting close~!

No comments:

Post a Comment

We Interrupt this Broadcast