Scribd
Upload
17 views

lec 3

Uploaded by

jsr.uk.coder
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
17 views

lec 3

Uploaded by

jsr.uk.coder
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 28

Machine Learning

COCSC403/CACSC403

Lecture 3:
Introduction of AI/ML/DL

Dr. Poonam Rani


Associate Professor
NSUT(CSE)
TOPICS

 Why students learn ML

 History of ML

 AI/ML/DL

 Types of ML

 BASIC CONCEPTS OF ML

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


History of Machine learning

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


History ….

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


What is Data?
Types of data

 Unstructured

 Structured

 Big data

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


Complexity in data and rules

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


Machine learning

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


DATA SCIENCE LIFE CYCLE
① DATA ACQUISTION
② DATA PREPROCESSING – 7 major Steps
① Getting dataset
② Importing libraries
③ Importing datasets
④ Finding missing values
⑤ Encoding categorical data
⑥ Splitting datasets
⑦ Feature scaling

③ MACHINE LEARNING ALGORITHM


④ PATTERN EVALUATION

⑤ KNOWLEDGE REPRSENTATION

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


How ML works

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014
Machine learning (ML) is a category of
an algorithm that allows software
applications to become more accurate
in predicting outcomes without being
explicitly programmed.

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


What is Machine Learning?
 Optimize a performance criterion using example
data or past experience.

 Role of Statistics: Inference from a sample

 Role of Computer science: Efficient algorithms to


 Solve the optimization problem
 Representing and evaluating the
model for inference

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


Types of Machine Learning
•– Linear
Supervised learning
regression
– Logistic regression
– Decision tree induction
– Support vector machines & kernel methods
– Model ensembles
– Bayesian learning
– Neural networks & deep learning
– Learning theory

• Unsupervised learning –
-Clustering
– Dimensionality reduction

• Reinforcementlearning
– Temporal difference learning
– Q/R/TD learning

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


Deep learning

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


Assignment 3.1
 Explore following applications
1. ML in computer vision
2. Duolingo mobile applications – Learning language
– French, Zerman like
3. Auto friend tagging suggestion
4. Recommendation

 10 Real life examples of each of AI/ML/DL

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014
AI

ML

DL

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


Artificial Intelligence
Artificial narrow intelligence (ANI)
Artificial general intelligence (AGI)
Artificial super intelligence (ASI)

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


ML vs DL
ML DL

Data Less data High data


Dependency

Hardware Simple as less Complex – GPU/ASIC


Dependency features

Feature Feature parameter Automatic feature selection


selection need to be given

Time in Training less Training more but


training and Testing more testing less
testing data

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014
Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014
23

Types of ML
 Types on the Level of ML

 Type of supervision

 Type of production environment

 Model based and instance based learning


Resources: Datasets
1. Kaggle

2. Amzon datasets

3. UCI Repository:
http://www.ics.uci.edu/~mlearn/MLRepository.ht
ml

4. Google data search engine

5. Scikit-learning datasets

6. Government datasets – us,uk,in

7. Statlib: http://lib.stat.cmu.edu/

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


Assignment 3.2
 What do you mean by hyper parameters ?

 What do you mean by explainable AI? Explain its


5 applications?

 Describe one and make video of that.

 Differentiate AI and Explainable AI

 One Case study of ML in group of two –


computer vision

 How own datasets can be formed?

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


Assignment 3.3
 All ML and DL libraries ppt as lecture 4

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014


Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014
Resources: Journals
 Journal of Machine Learning Research
www.jmlr.org

 Machine Learning

 Neural Computation

 Neural Networks

 IEEE Trans on Neural Networks and Learning


Systems

 IEEE Trans on Pattern Analysis and Machine


Intelligence

 Journals on Statistics/Data Mining/Signal


Processing/Natural Language
Processing/Bioinformatics/...

Dr. Poonam Rani - MACHINE LEARNING COCSC403/CACSC403 7/7/2014

You might also like