Scribd
Upload
39 views54 pages

Deep Learning M1

Deep Learning and AI PPTX

Uploaded by

niladrix719
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
39 views54 pages

Deep Learning M1

Deep Learning and AI PPTX

Uploaded by

niladrix719
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/ 54

Deep Learning

Artificial Intelligence
• AI is the new Electricity

• Electricity had once transformed countless industries:


transportation,
manufacturing, healthcare, communications, and more

• AI will now bring about an equally big transformation.


Machine Learning Basics
Machine learning is a field of computer science that gives computers the ability to
learn without being explicitly programmed

Machine Learning
Labeled Data algorithm

Training
Prediction

Data Learned model Prediction

Methods that can learn from and make predictions on data


Types of Learning
Supervised: Learning with a labeled training set
Example: email classification with already labeled emails

Unsupervised: Discover patterns in unlabeled data


Example: cluster similar documents based on text

Reinforcement learning: learn to act based on feedback/reward


Example: learn to play Go, reward: win or lose

class A

class A

Classification Regression Clustering

Anomaly Detection
Sequence labeling
http://mbjoseph.github.io/2013/11/27/measure.html
…
ML vs. Deep Learning
Most machine learning methods work well because of human-designed
representations and input features
ML becomes just optimizing weights to best make a final prediction
What is Deep Learning
(DL) ?
A machine learning subfield of learning representations of data. Exceptional effective
at learning patterns.
Deep learning algorithms attempt to learn (multiple levels of) representation by using
a hierarchy of multiple layers
If you provide the system tons of information, it begins to understand it and respond
in useful ways.

https://www.xenonstack.com/blog/static/public/uploads/media/machine-learning-vs-deep-learning.png
Application
• Real Estate
• Online Advertising
• Photo tagging
• Speech Recognition
• Machine Translation
• Autonomous Driving
Growth Rate
Why Deep Learning?
Growing Dataset Trend
Why is DL useful?
o Manually designed features are often over-specified, incomplete and take a long time
to design and validate
o Learned Features are easy to adapt, fast to learn
o Deep learning provides a very flexible, (almost?) universal, learnable framework for
representing world, visual and linguistic information.
o Can learn both unsupervised and supervised
o Effective end-to-end joint system learning
o Utilize large amounts of training data

In ~2010 DL started outperforming other


ML techniques
first in speech and vision, then NLP
Next Class
• Linear Algebra

• BooK-
• Deep Learning by
Ian Goodfellow, Yoshua Bengio and Aaron
Courville
• Publisher- MIT Press.
Linear Algebra
PEC 501
DEEP LEARNING
Focused on the subset most relevant to deep learn
Linear algebra and deep lear

 Linear algebra is central to the des


deep learning
 methods via matrix notation to the
implementation of deep learning m
such as Google’s TensorFlow Pytho
that has the word "tensor" in its na
Application

 The use if linear algebra structures when


data such as tabular datasets and image
 Linear algebra concepts when working w
preparation such as one hot encoding a
dimensionality reduction.
 The in-grained use of linear algebra nota
methods in subfields such as deep learn
language processing and recommender
Application of linear algebra

 1. Dataset and Data Files


 2. Images and Photographs
 3. One Hot Encoding
 4. Linear Regression
 5. Regularization
 6. Principal Component Analysis
 7. Singular-Value Decomposition
 8. Latent Semantic Analysis
 9. Recommender Systems
Recommender Systems

 Book recommendation by Amazon / recommendation


show by Netflix.

 Calculation of the similarity between sparse customer


using distance measures such as Euclidean distance o

 Matrix factorization methods like the singular-value de


used widely in recommender systems to extract item
their essence for querying and searching and compar
Topic of Todays discussion:

 scalars
 vectors
 Vector operations
 Vector norm
 Matrix Types
1.Square Matrix
2.Symmetric Matrix
3.Triangular Matrix
4.Diagonal Matrix
5.Identity Matrix
6.Orthogonal Matrix
Next Class

 Matrix Operation
 Sparse Matrix
 Tensor Arithmetic
 Factorization
 Matrix Decomposition
 Eigen Decomposition
 Singular Value Decomposition
Probability
PEC 501
DEEP LEARNING
Focused on the subset most relevant to deep learn
Linear algebra and deep lear
Probability Distribution

 relationship between observations and their probabil


 unimodal distribution, such as the familiar bell shape.
 Probability density function, or PDF (continuous proba
 Probability mass functions or pmf (discrete probability)
Difference between the Probability &

 Probability - finding the chance of something g


distribution of the data
 Likelihood - refers to finding the best distributio
given a particular value of some feature or som
data.
 Example

The likelihood in very simple terms means to incre


chances of a particular situation(maximum likeliho
happen/occur by varying the characteristics of the
distribution.
Maximum Likelihood Estim

 process of maximizing our desired value in a probabil


through estimation.
Maximum Likelihood Estim

 Our motive here is to maximize “mu” as it is shown in


normal distribution above, the value of the PDF is max
is maximum.
Maximum Likelihood Estim

 To find out the maximum value of “mu”, we hav


above-obtained derivative to 0 in order to maxi
After doing the same we will obtain:
Probability Distribution

 Statistical function that describes all the possible values and


random variable can take within a given range.
 Some outcomes of a random variable will have low pr
and other outcomes will have a high probability densi
Observations

 Multiple peaks that don’t disappear with different num


referred to as a bimodal distribution, or multiple pea
a multimodal distribution.
 You might also see a large spike in density for a given
range of values indicating outliers, often occurring o
distribution far away from the rest of the density.
 choose appropriate learning methods that require inp
specific probability distribution.

You might also like