/Machine Learning

Machine Learning Is Going Real-time

- Chip Huyen tl;dr: Chip discusses two approaches: (1) Online predictions, where an ML system makes predictions in real-time. (2) Online learning, where ML system incorporate new data and update models in real-time.

featured in #219


Experimenting With Automatic Video Creation From A Web Page

- Peggy Chi Irfan Essa tl;dr: "we envision a future where creators focus on making high-level decisions and an ML model interactively suggests detailed temporal and graphical edits for a final video creation on multiple platforms."

featured in #215


The Case For A Learned Sorting Algorithm

- Adrian Colyer tl;dr: On a large dataset i.e. 1 billion items, Learned Sort outperforms its competitor by a factor of 1.49x, and that includes time taken to train the model. Adrian explains how it works.

featured in #211


Optimal Peanut Butter And Banana Sandwiches

- Ethan Rosenthal tl;dr: "How do we make optimal peanut butter and banana sandwiches? You take a picture of your banana and bread, pass the image through a deep learning model to locate said items, do some nonlinear curve fitting to the banana, transform to polar coordinates and “slice” the banana along the fitted curve..."

featured in #202


How To Trick A Neural Network In Python 3

- Alvin Wan tl;dr: A tutorial that teaches you to trick an animal classifier and an understanding of how to defend against such tricks.

featured in #190


Face Detection Using JavaScript API - face-api.js

- Deepak Gupta tl;dr: Built on TensorFlow, this solves face detection, face recognition and face landmark detection, optimized for web and mobile devices.

featured in #157


Code As Craft: Understand The Role Of Style In E-commerce Shopping

- Aakash Sabharwal Jingyuan Zhou tl;dr: Overview of how Etsy built a ML tool that analyzes a user's taste and subsequently generates personalized recommendations.

featured in #150


Finding The Point Of Human Leverage

- Ben Evans tl;dr: Large platforms are mechanical turks relying on users to "create, capture and channel human annotation" e.g. FB knows your likes once you hit "like", in doing so it has found a point of leverage. Such platforms need a lot of users but, crucially, ML doesn't. This will change the point of leverage.

featured in #137


Bringing Black And White Photos To Life Using Colourise.sg 

- Preston Lim tl;dr: Using a deep learning technique called GANs, Colourise.sg predicts the color values at each pixel in black & white images. The model is trained on old Singaporean photos. In order to create accuracy, colorization requires historical, geographic, and cultural context. Below is an example, you can try the tool here on your own B&W photos.

featured in #132