How the Terminator could save everyone. You included.

The solution to all our problems? What? Source

Please welcome…the Terminator.

You heard (or read) me right.

How in Cyberspace does AI work?

Imagine a classroom, with a teacher that can test but not teach, and one student to look after. The student has only one metric that can tell him how he’s doing — his test score (since he literally wasn’t taught anything). His only goal is to increase his test scores and decrease mistakes.

Your brain is an AI. Kind of.

Wait what? What does my brain have to do with an AI?

I promise, they’re not as scary as they look. Source
This looks even scarier, but it’s just a bunch of perceptrons stacked on top of each other! Source
“Oh no — CALCULUS”
In this case, the error is the lowest when the weight is set to 2.

Some Problems…

The example we used is called an MLP (multilayer perceptron) — where each node is connected to each other node. They’re general purpose — trainable on text, audio, etc. — but aren’t really good at doing one task well.

Not the best handwriting, but at least the numbers are clear for the AI! Source
Check out 3Blue1Brown for an even better explanation on Neural Networks! Source

We need a Specialist.

While the name’s a mouthful, a CNN operates on basic concepts — and it’s all inspired by your eyes 👀.

The CNN architecture. This one uses the RELU activation function (which is more common than the sigmoid), and the fully connected layer is the MLP.
Max pool, and Average pool. Source

The Recurring Pain of Neural Networks.

RNNs are, by far one of the most powerful networks out there — all because they can remember things.

Yum 😋 Source
The hidden layer gets used as a memory, and is used the next time we forward propagate the RNN. Source

RNNs on Steroids (the last one, I swear)

Short for long-short-term-memory (try saying that three times fast), LSTMs are another special kind of neural network. They have 2 memories (long/short) and 4 gates to process them: learn, forget, update, and use.

LSTM Architecture — Source

Conclusion

You now probably agree that AI is pretty damn cool — but (like everything), it also has a couple problems. Mainly that it’s a little too good at achieving its goals, beating us at Go, Chess, and Jeopardy. It’s limited based on what we tell it do to and the goals we make.

--

--

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Aditya Dewan

Aditya Dewan

55 Followers

Building companies. Machine Learning Specialist @Actionable.co. Philosophy x Tech.