Tweet

@DilemmaBot

https://twitter.com/haltakov/status/1361439744018812929

@apoorv__tyagi

@AlejandroPiad

More from @haltakov

Vladimir Haltakov

@haltakov

17 Feb

What are Convolutional Neural Networks? 🏞️ ⏭️ ⛰️

CNNs are an important class of deep artificial neural networks that are particularly well suited for images.

If you want to learn the important concepts of CNNs and understand why they work so well, this thread is for you!

🧵👇

What is a CNN? 🤔

A CNN is a deep neural network that contains at least one convolutional layer. A typical CNN has a structure like this:
▪️ Image as input
▪️ Several convolutional layers
▪️ Several interleaved pooling layers
▪️ One/more fully connected layers

Example: AlexNet

https://twitter.com/haltakov/status/1310591421771059200

A good example - AlexNet

Throughout the thread I will be giving examples based on AlexNet - this is the net architecture that arguably started the whole deep learning revolution in computer vision!

I've written more about AlexNet here:

https://twitter.com/haltakov/status/1310591421771059200

Read 21 tweets

Vladimir Haltakov

@haltakov

15 Feb

Prisoner's Dilemma 🤔

Time for some game theory! 👨‍🏫

Prisoner's Dilemma (PD) is an interesting game that explains how two rational individuals may make decisions that seem irrational.

The game has lots of examples and applications in real life!

Thread 👇

There are different examples of PD, but this is the one I like most.

You want to buy something from another person. You exchange closed bags one containing the money and one the goods.

Both you and the other person can choose to honor the deal ✅ or to give an empty bag ❌.

If you both honor the deal ✅ ✅ (cooperate), you both gain something.

If you both exchange empty bags ❌ ❌ (defect), at least nobody loses.

If you leave the bag empty, but get a full bag ✅ ❌, you gain a lot, while the other person is screwed.

Image source: Wikipedia

Read 11 tweets

Vladimir Haltakov

@haltakov

10 Feb

Dealing with imbalanced datasets 🐁 ⚖️ 🐘

Real world datasets are often imbalanced - some of the classes appear much more often in your data than others.

The problem? You ML model will likely learn to only predict the dominant classes.

What can you do about it? 🤔

Thread 👇

Example 🚦

We will be dealing with a ML model to detect traffic lights for a self-driving car 🤖🚗

Traffic lights are small so you will have much more parts of the image that are not traffic lights.

Furthermore, yellow lights 🟡 are much rarer than green 🟢 or red 🔴.

The problem ⚡

Imagine we train a model to classify the color of the traffic light. A typical distribution will be:
🔴 - 56%
🟡 - 3%
🟢 - 41%

So, your model can get to 97% accuracy just by learning to distinguish red from green.

How can we deal with this? 🤔

Read 13 tweets

Vladimir Haltakov

@haltakov

8 Feb

Machine Learning Formulas Explained 👨‍🏫

This is the formula for Mean Squared Error (MSE) as defined in WikiPedia. It represents a very simple concept, but may not be easy to read if you are just starting with ML.

Read below and it will be a piece of cake! 🍰

Thread 👇

The core ⚫

Let's unpack from the inside out. MSE calculates how close are your model's predictions Ŷ to the ground truth labels Y. You want the error to go to 0.

If you are predicting house prices, the error could be the difference between the predicted and the actual price.

Why squared? 2️⃣

Subtracting the prediction from the label won't work. The error may be negative or positive, which is a problem when summing up samples.

You can take the absolute value or the square of the error. The square has the property that it punished bigger errors more.

Read 8 tweets

Vladimir Haltakov

@haltakov

28 Jan

Is this formula difficult? 🤔

This is the formula for Gradient Descent with Momentum as presented in Wikipedia.

It may look intimidating at first, but I promise you that by the end of this thread it will be easy to understand!

Thread 👇

The Basis ◻️

Let's break it down! The basis is this simple formula describing an iterative optimization method.

We have some weights (parameters) and we iteratively update them in some way to reach a goal.

Iterative methods are used when we cannot compute the solution directly

Gradient Decent Update 📉

We define a loss function describing how good our model is. We want to find the weights that minimize the loss (make the model better).

We compute the gradient of the loss and update the weights by a small amount (learning rate) against the gradient.

Read 7 tweets

Vladimir Haltakov

@haltakov

27 Jan

How to add new classes to your ML model? 🍏🍎🍊... 🍌?

You have a large multi-class NN in production.

You discover a new important class and want to add support for it *quickly* and with *low* risk.

Example: traffic signs recognition for self-driving cars 🛑🚗

Thread 👇

The naive approach 🤷‍♂️

Collect examples of the new class (for example a new traffic sign), label them and retrain the whole NN.

✅ It will probably work

❌ It will be time consuming, especially for big models.
❌ Risk for unintended regressions

Freezing the first layers 🥶

Typical CNNs learn generic image features in the initial layers and they will likely apply to the new sign as well.

You can freeze the weights of the initial layers and only retrain the last fully connected layer(s).

Read 10 tweets

Share this page!

Vladimir Haltakov

Try unrolling a thread yourself!

More from @haltakov

Vladimir Haltakov

Vladimir Haltakov

Vladimir Haltakov

Vladimir Haltakov

Vladimir Haltakov

Vladimir Haltakov

Did Thread Reader help you today?

Like this author's thread?