Delta Academy

Live courses to learn AI.

Learn from Reinforcement Learning from zero-to-AlphaGo with weekly AI-building competitions.

[Delta Academy @Delta_Academy_]

Our next Introduction to Reinforcement Learning cohort is live on Monday!

Join us at the link below to learn to build AIs through live games

https://t.co/nZReAvtVNa

[Mario Kostelac @mariokostelac]

Finally found a gap in schedule to join @Delta_Academy_.

Let's have some fun and learn about Reinforcement Learning.

If you're interested in doing the same, reply. I'm happy to share my feedback as I go. It lasts for 4 weeks.

[Delta Academy @Delta_Academy_]

A new Intermediate Reinforcement Learning course kicks off this week!

There's still time to join - you'll learn to build AIs using algorithms up to AlphaGo 🚀

For a taster, check out the Tron AI below, built by a student using Monte Carlo Tree Search

https://t.co/IoSKuoiuRd

[Delta Academy @Delta_Academy_]

The boom in AI we are seeing was kicked off by DeepMind's DQN algorithm, which conquered Atari games.

But it turns out, it's pretty tricky to train

We dig into why and give you the tricks that were needed to get it to work

https://t.co/YJhFgMIFrh

[Delta Academy @Delta_Academy_]

You may have heard that you have a learning style.

Well, it turns out there is little concrete evidence that you do - it's probably a myth

We dig into why this myth spreads, and what it shows about how we think about learning

https://t.co/SosaCDTitN

[Delta Academy @Delta_Academy_]

Trying to learn a new skills?

Though it's counterintuitive, it's best to dive right in the deep end.

Our co-founder @Phillips_M_G outlines why, and how Delta Academy is built to capture the best of doing so

https://t.co/BVhVcEqBtP

[Delta Academy @Delta_Academy_]

What if we could speed up learning?

It turns out, we can. A project by the US DARPA program showed we can increase learning rate by 2 standard deviations using a digital tutor

We dig into the study and it's implications

https://t.co/6eqn2KcWtk

[Delta Academy @Delta_Academy_]

When learning a new skill, it's best to dive in deep

We're building 'Immersive learning experiences' - the most effective way to learn skills

https://t.co/BVhVcEIcln

[Delta Academy @Delta_Academy_]

it's just true

[Delta Academy @Delta_Academy_]

@terrymac51 Hey, this is just where the instructors are from!

Other universities have experts too

[Delta Academy @Delta_Academy_]

Learn Reinforcement Learning live with Delta Academy.

Each week, build an AI that competes live against peers, guided by experts from Oxford, Cambridge and DeepMind.

Join our next 4 week cohort today.

[Delta Academy @Delta_Academy_]

Just another reason to learn RL

[Google DeepMind @GoogleDeepMind]

Today in @Nature: #AlphaTensor, an AI system for discovering novel, efficient, and exact algorithms for matrix multiplication - a building block of modern computations. AlphaTensor finds faster algorithms for many matrix sizes: https://t.co/E18DezRPTL & https://t.co/SvHgsa0SNV 1/

[Mike Schroepfer @schrep]

Maybe my favorite AI generated video so far. Prompt: "A fluffy baby sloth with an orange knitted hat trying to figure out a laptop, close up, highly detailed, studio lighting, screen reflecting in its eye.mp4"
killer work @MetaAI !!

[Delta Academy @Delta_Academy_]

And here's team Anudeep, who wraps up our finalists this week

Congratulations to all our competitors!

https://t.co/dSee8iqG2g

[Delta Academy @Delta_Academy_]

Our latest Introduction to Reinforcement Learning competition: Connect 4!

The teams built AIs using feature vectors

Congratulations to Team Sam on their victory 🎉

You can play against the different AIs in the thread below 👇🧵

https://t.co/WPH8AcGRIE

[Delta Academy @Delta_Academy_]

Our third placed finisher, Team Blue Ice Cream

Here's the link to play against their AI:

https://t.co/vF2jvyncon

[Delta Academy @Delta_Academy_]

Congratulations to this weeks Wild Tic Tac Toe Champion - Team Sam!

Play against their AI here:

https://t.co/wLlHFRDiF2

[Delta Academy @Delta_Academy_]

What a finale to the Tron competition tonight!

The teams were implementing Monte Carlo Tree Search - a method for planning.

Check out the Blue teams finishing move 😂🔥

https://t.co/j8LwItmRN1

[Delta Academy @Delta_Academy_]

If you want to dive deeper, follow to stay in touch with future threads.

You can also join a Delta Academy 4 week course. Each week you build an AI that competes live against a cohort in iconic games

Our intro course goes from 0-to-DQN, join here:

https://t.co/ekaF5Xpa6b

[Delta Academy @Delta_Academy_]

Reinforcement Learning can seem daunting - but it's a really powerful way to build AIs

It's way easier with short tutorials and interactive exercises in #Python

Let's go through the 1st of our Introduction to Reinforcement Learning (RL) tutorials, tweet-by-tweet 🧵👇

[Delta Academy @Delta_Academy_]

Conclusions:

RL solves sequential decision problems. They are set up as MDPs, with states, actions and rewards.

It aims to find the optimal policy

In the next thread, we'll start to look at the representations for finding this optimal policy

https://t.co/Gnm7ld6M29