Arxiv Insights

Arxiv Insights is a channel about artificial intelligence made by Xander Steenbrugge. He has worked in AI since 2012, and shares ideas from research and from his own projects.

The best known videos explain research methods in machine learning, such as reinforcement learning and autoencoders, using diagrams and clear narration. They are more technical than most science channels, so they suit students who are already comfortable with maths. More recent uploads look at AI tools for art and animation, and at how AI is changing science.

Some highlights include:

An introduction to reinforcement learning, the method that lets an AI learn by trial and reward.

An explanation of variational autoencoders, a type of neural network that can learn to generate new images.

How the AlphaGo system learned to play the board game Go at a world class level.

Arxiv Insights has over 100,000 subscribers.

AI is now part of almost every branch of engineering, from self driving cars to designing new materials. These videos give a real sense of how the methods work, not just what they can do. They are a good way into computer science and software engineering, and into robotics, where reinforcement learning is widely used. Students who enjoy these could try the free coding tools mentioned in the videos to train a simple model themselves.