Reinforcement vs Supervised Learning

A supervised learning algorithm is used when we have a labeled training dataset.

Reinforcement learning is used in a scenario where an agent interacts with an environment.

to observe a basic behaviour and change its state to maximize its rewards or goal.

Reinforcement learning is different from supervised learning, the kind of learning studied in most current research in the field of machine learning. Supervised learning is learning from a training set of labeled examples provided by a knowledgable external supervisor. Each example is a description of a situation together with a specification — the label — of the correct action the system should take to that situation, which is often to identify a category to which the situation belongs. The object of this kind of learning is for the system to extrapolate, or generalize, its responses so that it acts correctly in situations not present in the training set. This is an important kind of learning, but alone it is not adequate for learning from interaction. In interactive problems it is often impractical to obtain examples of desired behavior that are both correct and representative of all the situations in which the agent has to act. In uncharted territory — where one would expect learning to be most beneficial — an agent must be able to learn from its own experience

The Q-learning algorithm is most used as a basic reinforcement algorithm. It does not require a model of the environment (hence “model-free”), and it can handle problems with stochastic transitions and rewards without requiring adaptations. The “Q” in Q-learning stands for quality, which represents how valuable the action is in maximizing future rewards. Moreover, It uses the environment rewards to learn over time the best action to take in a given state. In the below implementation, we have our reward table “P” from where the agent will learn. Using the reward table it chooses the next action if it’s beneficial or not and then updates a new value called the Q-value. The newly created table is called the Q-table and maps to a combination called (state, action) combination. If the Q-values are better, we have more optimized rewards.

For example, In the environment, there are four possible locations where you can drop off the passengers: R, G, Y, B or [(0,0), (0,4), (4,0), (4,3)]in (row, col) coordinates if you can interpret the above-rendered environment as a coordinate axis. If the taxi is faced with a state that includes a passenger at its current location, it is highly likely that the Q-value for pickup is higher when compared to other actions, like drop off or north.

Q-values are initialized to an arbitrary value, and as the agent exposes itself to the environment and receives different rewards by executing different actions, the Q-values are updated using the equation:

An example code below.

On the other hand,

Supervised learning is a machine learning task where an algorithm is trained to find patterns using a dataset. The supervised learning algorithm uses this training to make input-output inferences on future datasets. In the same way a teacher (supervisor) would give a student homework to learn and grow knowledge, supervised learning gives algorithms datasets so it too can learn and make inferences.

The ultimate goal of the supervised learning algorithm is to predict Y with the maximum accuracy for a given new input X. There are several ways to implement supervised learning and we’ll explore some of the most commonly used approaches.

Based on the given data sets, the machine learning problem is categorized into two types: classification and regression. If the given data has both input (training) values and output (target) values, then it is a classification problem. If the dataset has continuous numerical values of attributes without any target labels, then it is a regression problem.

References:

Page 2, Reinforcement Learning, Richard S. Sutton & Andrew G. Barto

Chapter10, Hands-On Exploratory Data Analysis with Python

by Suresh Kumar Mukhiya, Usman Ahmed

Related posts:

Cosmic JS makes it easy to manage content for your Bootstrap applications. In this blog we’ll quickstart a Bootstrap Landing Page using the Cosmic CLI. This single page website landing page is built…

2020 is going to be the year of technological start-ups, and companies related to cryptocurrency, Big Data, AI, and other related fields are ready to take the stage. As common with the various…

Have you ever wondered why some people advance rapidly in their careers and others don’t? Would you like to climb the ladder of success but don’t know how? Here are 14 traits of people who achieve…