# An intro to Reinforcement Learning (with otters)

DevFeed: [An intro to Reinforcement Learning (with otters)](<https://devfeed.tech/articles/an-intro-to-reinforcement-learning-with-otters-35541.md>)

Original publisher: [Read original article](<https://meowni.ca/posts/rl-with-otters/>)

Author: Monica Dinculescu

Published: 2018-02-26T00:00:00Z

Content type: tutorial

Language: en

Sources: [Monica Dinculescu](<https://devfeed.tech/sources/monica-dinculescu.md>)

Topics: [Reinforcement learning](<https://devfeed.tech/topics/reinforcement-learning.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Learning](<https://devfeed.tech/topics/learning.md>)

Tags: [learning](<https://devfeed.tech/tags/learning.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [reinforcement-learning](<https://devfeed.tech/tags/reinforcement-learning.md>)

## AI overview

A beginner-friendly introduction to reinforcement learning explains how agents learn to choose sequences of actions to maximize long-term rewards. It contrasts reinforcement learning with supervised and unsupervised learning and uses accessible examples, including a Roomba and otter image classification.

## Source excerpt

Before I wrote the JavaScripts, I got a master's in AI (almost a decade ago 🙀), and wrote a thesis on a weird and new area in Reinforcement Learning. Or at least it was new then. It's definitely still weird now. Anyway, I loved it. With all the hype around Machine Learning and Deep Learning, I thought it would be neat if I wrote a little primer on what Reinforcement Learning really means, and why it's different than just another neural net. Richard Sutton and Andrew Barto wrote an amazing book called "Reinforcement Learning: an introduction"; it's my favourite non-fiction book I have ever read in my life, and it's why I fell in love with RL. The complete draft is available for free here, and if you're into math, and want to explore this topic further, I can't recommend it enough. If you're not into math, I have otters. What is it? Reinforcement learning (or RL) solves a very specific problem: figuring out how to act over time, so that you get the most long term reward. Both these sequences of actions and the reward bit are important components that make RL a "good" approach to solve a problem. For example, this is perfect if you're a Roomba who is trying to get home (the only reward you get is if you actually get home, so while you're roaming around aimlessly and get no 💰, you have a feeeeeeeling you're not doing it right). On the other hand, this is terrible if you're trying to figure out if a photo has an otter in it; there are no sequences of actions that matter here, other than doing the decision of saying "yes iz otter". You're just trapped in a room where people slip Polaroids of animals under the door and you have to tell them what it is. Nightmares aren't really a good area for RL. What isn't it? There are many things with the word "learning" in them that aren't Reinforcement Learning. supervised learning. This is a kind of Machine Learning where someone gave you a training set that has everything labelled correctly, you learn from it, and hope that at exam