AI outperforms humans in many tasks. As AI is used more widely, it will be important to understand how algorithms make decisions. Observable has notebooks for aspiring ML students and professionals.

Sangwon Jeong

Interactively Assessing Disentanglement in GANs

denis parra

PUC Chile

Analyzing the Design Space for Visualizing Neural Attention in Text Classification

Jeya Balaji Balasubramanian

Tensors

Image tensors

Kai Middlebrook

Visualization in Deep Learning

Thomas Preusse

Interactive Convolutional Neural Network

"Ta" Virot C.

Machine Learning in JS with TensorFlow.js (Part I)

WinkJS

How to visualize timeline of a Wiki article?

Payton Soicher

Simulation Results

Ryan Seddon

Machine learning in the browser

Sandraviz

Recommender systems

Kreijstal

Python in Observable through Pyodide

Giuliano Giacaglia

Artificial intelligence systems Timeline

Radamés Ajna

background-position Scrubber

Hugging Face

Drawings to Human

Sylvain Lesage

sahajBERT - Bubbles chart

Pretrained models (image classification)

Advanced pretrained models (object detection)

Justin Jay Wang

OpenAI

Robotics Perturbation: Broken Joint (Rotation)

Multi-Agent Evaluation: Object Counting

Observable

Collections related to specific areas of interest, curated by the Observable team. 

Topics

Artificial Intelligence

A collection of notebooks exploring TensorFlow.js

TensorFlow.js

Probably good to mention that tf.tidy cannot take in an async function, and the return value cannot be a Promise : https://js.tensorflow.org/api/latest/#tidy .

Praful Bhawsar

Calculation and visualization for the coefficients can be seen at:
https://observablehq.com/d/907eea004b9a540a

Mario Delgado 

Proof of the logit can be found here: 

https://observablehq.com/d/f29b4d8b1a4acf9b?collection=@mariodelgadosr/youre-never-too-old-to-code#cell-1169

Perhaps more readable to do `tf.tensor(myTensor.arraySync()).buffer()` ?