[go: up one dir, main page]
Scribd
Upload
17 views17 pages

Lecture1 Introduction Part1

The document provides an overview of reinforcement learning (RL) and its distinction from other machine learning branches, such as supervised and unsupervised learning. It emphasizes the unique characteristics of RL, including the absence of a supervisor, delayed feedback, and the impact of an agent's actions on future states. Additionally, it discusses the applications of RL in various fields, including gaming, robotics, and autonomous systems.

Uploaded by

mscai2024.avinesh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
17 views17 pages

Lecture1 Introduction Part1

The document provides an overview of reinforcement learning (RL) and its distinction from other machine learning branches, such as supervised and unsupervised learning. It emphasizes the unique characteristics of RL, including the absence of a supervisor, delayed feedback, and the impact of an agent's actions on future states. Additionally, it discusses the applications of RL in various fields, including gaming, robotics, and autonomous systems.

Uploaded by

mscai2024.avinesh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 17

Introduction to Reinforcement

Learning
Revolution History
Branches of ML
Branches of ML - Supervised Learning
• In Supervised Learning, models learn from labeled
training data, where input-output pairs are provided.

• The algorithm generalizes from this labeled data to


make predictions or classifications on new, unseen
data.

• Commonly used in tasks like image recognition,


natural language processing, and regression analysis.
Branches of ML - Unsupervised Learning
• UL deals with unlabeled data, aiming to discover
patterns, structures, or relationships within the data
itself.

• Clustering and dimensionality reduction are common


tasks in UL.

• Applications include customer segmentation, anomaly


detection, and feature extraction.
Branches of ML - Reinforcement Learning
• RL involves an agent learning to make decisions by
interacting with an environment.

• It receives feedback in the form of rewards or


penalties, guiding the agent toward optimal decision-
making strategies.

• RL is well-suited for scenarios where actions influence


future states, making it applicable in gaming, robotics,
and autonomous systems.
Can Machines Think?

The imitation game – Movie

Computing Machinery & Intelligence – Paper


What is intelligence according to
you?

-To be able to make decisions to achieve the goal


What is RL?
Example

Learning by interacting with the environment


RL Characteristics
• What makes reinforcement learning different
from other machine learning paradigms?
– There is no supervisor, only a reward signal
– Feedback is delayed, not instantaneous
– Time really matters - sequential
– Agent’s actions affect the subsequent data it
receives
Agent Environment Loop
Reward Hypothesis
• Any goal can be formalized as the outcome of
maximizing a cumulative reward

• Also we can consider minimizing the penalty


RL Problems
• Fly helicopter – inverse distance

• Walking robot – distance, speed

• Board games - maximize score or +1 (win) -1


(lose)
Reasons to learn
• Find a solution
– A program that plays chess very well
– A manufacturing robot with a specific purpose

• Adapt online to handle unforeseen


circumstances
– Chess program can learn to adapt to you
– Candy crush
– A robot that learns to navigate unknown terrains
What is RL?

Science and framework to make decisions from interactions


Thank You

You might also like