Scribd
Upload
14 views18 pages

4 5 TypesOfLearning

The document discusses four main types of machine learning: supervised learning, unsupervised learning, reinforcement learning, and semi-supervised learning. Supervised learning uses labeled training data to learn a function to map inputs to outputs. Unsupervised learning finds hidden patterns in unlabeled data. Reinforcement learning involves an agent learning through trial-and-error interactions with an environment. Semi-supervised learning uses a small amount of labeled data and a large amount of unlabeled data.

Uploaded by

Himanshu Chhabra
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
14 views18 pages

4 5 TypesOfLearning

The document discusses four main types of machine learning: supervised learning, unsupervised learning, reinforcement learning, and semi-supervised learning. Supervised learning uses labeled training data to learn a function to map inputs to outputs. Unsupervised learning finds hidden patterns in unlabeled data. Reinforcement learning involves an agent learning through trial-and-error interactions with an environment. Semi-supervised learning uses a small amount of labeled data and a large amount of unlabeled data.

Uploaded by

Himanshu Chhabra
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 18

Types of Machine Learning

There are four different types of M/L:


Supervised Learning

Unsupervised Learning

Reinforcement Learning

Semi Supervised Learning


Supervised Learning

Labelled data New Input X


Direct feedback X Y
Input 1 Output 1
Predict outcome / Input 2
Input 3
Output 2
Output 3
Learning
Model
Model
future .
.
.
.
Input n Output n

Output Y

Given a set of pairs (x1, y1), (x2, y2), …, (xk, yk)


Learn a function f(x) to predict y for (unseen) x
Supervised Learning
 Data labeled by human experts
 X,Y (pre- classified training examples)
 Given an observation x, what is the best
label for y?
 Labeling images
 Speech recognition
 OCR (Optical Character Reorganization)
Supervised Learning

Regression
 Output is continuous (real-valued).
 Predicting price of a house, given its location, carpet area,
number of rooms, …
Classification
 Output is discrete (integer-valued / classes / labels).
Given the weather for last 7 days, predict whether it will rain
tomorrow.
Classification

Example: Credit scoring


Differentiating between
low-risk and high-risk
customers from their
income and savings

Discriminant: IF income > θ1 AND savings > θ2


THEN low-risk ELSE high-risk
Classification: Applications
Pattern recognition
Face recognition: Pose, lighting, occlusion (glasses, beard), make-up,
hair style
Character recognition: Different handwriting styles.
Speech recognition: Temporal dependency.
Medical diagnosis: From symptoms to illnesses
Biometrics: Recognition/authentication using physical and/or
behavioral characteristics: Face, iris, signature, etc
Outlier/novelty detection:
Regression

 Example: Price of a used


car
 x : car attributes
y : price
y = g (x | q )
g ( ) model,
q parameters
Supervised Learning
Supervised Learning
Unsupervised Learning

Image: https://www.toppr.com/content/story/amp/temperature-as-measure-of-hotness-or-coldness-105568/
Unsupervised Learning

 Unlabeled data X.
 Given a set of x’s, cluster or
summarize them.
 Grouping customers
 Detecting new diseases
 Anomaly detection
Unsupervised Learning
Clusters
No labels
X Inputs 1,7,8, …
No feedback Input 1
Input 2 Learning Inputs 2,11,…
Find hidden Input 3
.
Algorithm

structure in data .
Input n
Inputs 9, 12, …

Given a set of elements x1, x2, …, xk,


Learn some hidden structure/organisation
Reinforcement Learning

Image: https://towardsdatascience.com/reinforcement-learning-the-naturalist-the-hedonist-and-the-disciplined-fc335ac9289c
Reinforcement Learning

Decision process
Reward system
Learn a series of actions
Trial and error

From a sequence of rewards/penalties for actions


Learn how to maximize the rewards (minimize penalties)
Reinforcement Learning

Decision process
Reward system
Learn a series of actions
Trial and error
Performance
Agent

Environment Knowledge

Environment
Learning

Ref: Dietterich & Langley, 2003


Semi Supervised Learning

Semi-supervised machine learning is a combination of


supervised and unsupervised machine learning methods.
 Learning from a small set of labelled data
Training set consists of a small amount of labeled data
along with a large amount of unlabeled data
Used in Speech Analysis, Image Analysis, Protein Sequence
Classification
Semi Supervised Learning
Other types of learning approaches
Generative approaches
Creating new data that is “like” given data.
Generally included in unsupervised learning
Eg Out of 1000 cat images design a new cat image
having only key features of cat image.

Self-supervised learning
Implicit labels are extracted from data using
heuristics.
Eg Predict new video frame. Predict some missing data
based on given data set.

You might also like