Fundamentals

Applications of
Artificial Intelligence
(ME3181)
Fundamentals
Phung Thanh Huy

Department of Mechatronics
Ho Chi Minh City University of Technology (HCMUT)
huypt@hcmut.edu.vn
05/09/2023
Modelling from data
Well-posed Learning problem
A computer program is said to learn from experience E with respect to some
task T and some performance measure P, if its performance on T, as measured
by P, improves with experience E.
—Tom Mitchell, 1997
Lecture notes of Andrew Ng

Applications of AI (ME3181) 2
Modelling from data
Handwriting Recognition
Task T
–recognizing and classifying handwritten
words within images
Performance measure P
–percent of words correctly classified
Training experience E
–a database of handwritten words with
given classifications
Lecture notes of Andrew Ng

Modelling from data
Traditional Programming
Input
Computer Output
Rule
Machine Learning
• Training Phase / Learning Phase/ Modelling Modelling Process
Known Input
Computer Rule
Known Output
Model
Data
- Set of available data pieces = dataset
- Each data piece (point) is a sample or an example
- Data could be only input (unsupvised learning)
• Prediction Phase
New Input
Computer Predicted Output
Model
Modelling from data
o Each feature can be represented in a variable 𝑥𝑖
o All the features can be represented in a vector of features:
𝑥1
𝑥2
𝑥 = … = 𝑥1 , 𝑥2 , … , 𝑥𝑛 𝑇
𝑥𝑛
o A label can be represented in the same ways.
o There are some special ways to represent labels and features for some
special problems
Modelling from data
• Inputs: Features, Variables, Attributes, Covariates
• Outputs: Response variables, Labels, Tagets
https://www.amybergquist.com/
Modelling from data
Modelling from data
Modelling from data
Modelling from data
Features
Modelling from data
https://www.linkedin.com/pulse/customer-segmentation-using-python-data-science-deepak-sharma
Modelling from data
Camelia Florea et al., Bul. Inst. Polit. Iaşi, t.

LVII (LXI), f. 6, 2011
Modelling from data
o Supervised Learning
o Dataset is the collections of labeled examples.
o Un-supervised Learning
o Dataset is a collection of un-labeled examples.
o Self-supervised Learning (Semi-Supervised Learning)
o Dataset included both labled and un-labled examples.
o The number of un-labled examples is usually more than that of labled
examples
Modelling from data
examples
Modelling from data
examples
Modelling from data
examples
Modelling from data
https://www.analyticsvidhya.com/
Modelling from data
o Reinforcement Learning
o Learn a policy
Modelling from data
o Batch Learning (Offline Learning)
• Trained using all the available data
• First the system is trained, and then it is launched into production and
runs without learning anymore
• Take a lot of time and computing resources
o Online Learning (Incremental learning)
• Data is feed instances sequentially, either individually or in small
groups called minibatches
Using online learning to handle huge datasets
Modelling from data
o Instance-based Learning
• The system “learns the examples by heart”
• Then generalizes to new cases by using a similarity measure to
compare them to the learned examples (or a subset of them)
o Model-based Learning
• Build a model of the examples and the model to make
predictions
Data Preparation
Test = Val
Train Val Test
Data Preparation
• Popular open data repositories
UC Irvine Machine Learning Repository
Kaggle datasets
Amazon’s AWS datasets
• Meta portals (they list open data repositories)
Data Portals
OpenDataMonitor
Quandl
• Other pages listing many popular open data repositories
Wikipedia’s list of Machine Learning datasets
Quora.com
The datasets subreddit
Mathematical Concepts Review 1
Data structure
• Scalar: a simple numerical value.
• E.g.: 0, 1, 3.5, 2.5,…
• Scalar Variables are usually denoted as italic letters (𝑥, 𝑎, 𝑦, …).
• Vector (programing: array): ordered list of scalar values.
• The scalar values in a vectors are attributes or features.
• Usually denoted as bold letters x or w.
• Usually represented as column vectors.
• Lazy notation (in my lectures) : 𝑥 = 𝑥1 , 𝑥2 , … , 𝑥𝑛
• Can be represented as points or arrows
• Matrix: 2-D vector (lazy thinking)
• Tensor: Higher dimension vectors (lazy thinking)
Operations
• Summation
• Production
• Operations on Vector: Self-review (Norm of a vector)
• Operations on Matrix: Self-review
Function
Extrema: Minima, maxima
Derivative and Gradient
A function 𝑓(𝑥)
𝑑𝑓
Derivative: 𝑓 ′ 𝑥 = 𝑑𝑥
𝜕𝑓
Partial derivatives (multi-variable function): 𝜕𝑥
1
𝜕𝑓 𝜕𝑓
Gradient: ∇𝑓 is a vector [ , ,…]
𝜕𝑥1 𝜕𝑥2
Chain rules:
Random Variables and Probability
Source: UST Class-Machine Learning (by Junseong Bang)

Fundamentals

Uploaded by

Copyright:

Available Formats

Fundamentals

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fundamentals

Uploaded by

Copyright:

Available Formats

Applications of

Phung Thanh Huy

Lecture notes of Andrew Ng

Lecture notes of Andrew Ng

Camelia Florea et al., Bul. Inst. Polit. Iaşi, t.

Using online learning to handle huge datasets

Train Val Test

• Operations on Vector: Self-review (Norm of a vector)

• Operations on Matrix: Self-review

Source: UST Class-Machine Learning (by Junseong Bang)

Source: UST Class-Machine Learning (by Junseong Bang)

Source: UST Class-Machine Learning (by Junseong Bang)

Source: UST Class-Machine Learning (by Junseong Bang)

Source: UST Class-Machine Learning (by Junseong Bang)

Source: UST Class-Machine Learning (by Junseong Bang)

You might also like