How machine learning works and why it is such a job killer
We all heard about machine learning but how many of us actually know how that is done?
Here is a video that was linked by someone earlier that is fascinating. It really shows how machines can learn. It is about 13 minutes but you can fast forward past some of the repetitive stuff.
The reason why this is important is because it shows the capabilities of these systems. This is just a game so it is really a bit childish yet it is something that can be transferred to so much bigger. This system was training computerized ninjas, others are training for self driving cars or how to make a burger.
One thing to keep in mind is that this program can be run millions of times until the learning is complete. This is why simulation is so effective. Waymo, the self driving car company started by Google, has driven billions of miles in simulation.
The narrow nature of the "intelligence" makes it a slow process to create algorithms to handle a complete job. However, as we see, it is fairly easy to train a computer for a specific task. Start linking the tasks together and you will soon have AI that can handle most of the job.
It seems AI these days is based upon raw processing power and the ability to run millions of simulations until it learns what it needs to. With the advent of so much data, we now have a ton of information to feed these systems.
Something as simple as this video shows the power of what is being developed. What will the capabilities be in another 5 years?
That is what everyone needs to ask themselves.
To get your weekly distribution of manna, sign up here: https://www.mannabase.com.
For more information about the Manna Project and how you can contribute, visit https://www.peoplescurrency.org/
Unpaid related articles
Related articles
Interesteem(@interesteem) is a service that recommends related articles using DeepLearning.
Please write an article with #interesteem tag.
Recent related articles
Older related articles
Interesteem(@interesteem) is a service that recommends related articles using DeepLearning.
Please write an article with #interesteem tag.