In this edition you'll see 2 different experiments that I've done. The goal of these experiments is to make difficult concepts in machine learning easier to learn.

I used ChatGPT to not only simplify machine learning concepts, but to also explain them in a tone of a celebrity!

Experiment 1 : Shaan Puri explains 3D convolution to you.

Experiment 2 : Kevin Hart explains Tensorflow Extended to you.

Enjoy!

Shaan from MyFirstMillion podcast explains 3D convolution

​

I love machine learning and I love MyFirstMillion podcast (thanks Sam Parr and Shaan Puri!), so I thought of this small experiment.
​
Can ChatGPT rewrite a technical concept (3D convolution in CNNs) with the tone of Shaan Puri? Here's the result ↓
______________________

"Listen up folks, I'm about to drop some knowledge on you about 3D Convolution.
​
This bad boy is taking convolution to the next level by not just considering the height and width of an image or video, but also the DEPTH or TEMPORAL information.
​
That's right, we're talking about three dimensions, baby! And you know what they say, the more dimensions, the merrier!
​
Now, unlike 2D Convolution where the kernel is just sliding through the data in two dimensions, the 3D Convolution kernel is moving and grooving through the data in all three dimensions.
​
This means that the model is getting a deeper understanding of the data, and that's a good thing, folks.
​
So, when do we use this technique?
​
Well, let me tell you, it's a game-changer in medical imaging. We're talking about CT scans, MRI scans, and the like.
​
The model needs to understand the 3D structure of the image, and that's where the 3D Convolution comes in, making sure that the predictions are accurate and the performance is top-notch.
​
And let's not forget about video data. It's got a temporal dimension, and that makes it perfect for 3D Convolution to learn the spatiotemporal features of the data.
​
In conclusion, folks, 3D Convolution is where it's at. It's taking convolution to new heights, and that's what we like to see in this world of machine learning."
______________________
If you ever listened to the mfm podcast, then you'd know that the tone is very much that of Shaan Puri 😁

Kevin Hart explains Tensorflow Extended (TFX)

Yo, listen up! I want to talk about TFX, the production ML pipeline's BFF. When you're ready to take your models from the sandbox to the big leagues, TFX is the way to go!
​
With TFX, you can create a pipeline that will run like a well-oiled machine, making sure that your models are ready for the red carpet.
​
And let me tell you, these pipelines are a sequence of components that are specifically designed to handle large-scale, high-performance machine learning tasks. That's like the superhero team-up of machine learning!
​
Now, the best part? You can use TFX libraries like TensorFlow Data Validation, TensorFlow Transform, and TensorFlow Model Analysis to build your components.
​
Or, if you're feeling adventurous, you can write your own custom code.
​
And if you're working with a big team or processing a ton of data, TFX can even distribute the processing across a compute cluster using Apache Beam.
​
That's like the cavalry riding in to save the day!
​
In conclusion, TFX is the key to unlocking the full potential of your production ML pipeline.
​
Don't be left behind, get on the TFX train and experience the ease and efficiency of a fully automated pipeline.
​
Just remember, with great power comes great responsibility... and a lot of data to process!
​
More on Tensorflow Extended can be found here.