SlideShare a Scribd company logo
Machine Learning
Best practices and Vulnerabilities
MACHINE LEARNING
BEST PRACTICES AND VULNERABILITIES
Sebastiano Galazzo
Microsoft MVP
@galazzoseba
sebastiano.galazzo@gmail.com
https://it.linkedin.com/in/sebastianogalazzo
Best practices
THE PERCEPTRON
WPC2017 4
In machine learning, the perceptron is a binary classifier or a function which can decide whether or
not an input, represented by a vector of numbers, belongs to some specific class.
𝑓 𝑥 = χ ( ⟨ w , x ⟩ + b )
w is a vector having weights of real values, while operator ⟨ ⋅ , ⋅ ⟩ is the scalar product, b is the 'bias’,
a constant not related to any input value and χ ( y ) is the output function
MAIN EVOLUTIONS
1. Easy way, Logistic Regression, Support Vector Machine
• Pro: Easy and fast use use
• Cons: Low accuracy (compared to neural networks)
2. Hard way, Neural Networks
• Pro: If get convergence you gain a very high accuracy (State of the art)
• Cons: Very difficult to model, a lot of experience is required
WPC2017 5
EASY WAY
• Pseudo equation
• 𝑥 ∗∝ +𝑦 ∗ 𝛽 + 𝑐 ∗ 𝛿+. . +𝑧 ∗ 𝜔 = (0,1)
• #logisticregression #svm
WPC2017 6
HARD WAY
• #neuralnetwork
WPC2017 7
ADVANCED MODELLING OF NEURAL NETWORKS
WPC2017 8
Use case, provide a customer's willingness to vote a political party
Age Gender Income City Political party
30 Male 38,000 New York Democrat
39 Female 42,000 Page Republican
24 Other 39,000 San Francisco Republican
51 Prefer not to say 71,000 Seattle Democrat
ADVANCED MODELLING OF NEURAL NETWORKS
WPC2017 9
> 20 parameters
0,17 18,24 25,35 36,45 46,60 > 60 𝑚𝑎𝑙𝑒 𝑓𝑒𝑚𝑎𝑙𝑒 𝑢𝑟𝑏𝑎𝑛 𝑟𝑢𝑟𝑎𝑙 [𝑠𝑢𝑏𝑢𝑟𝑏𝑎𝑛]..[democrat][Republican]
Age Gender Income City Political party
30 Male 38,000 New York Democrat
39 Female 42,000 Page Republican
24 Other 39,000 San Francisco Republican
51 Prefer not to say 71,000 Seattle Democrat
ADVANCED MODELLING OF NEURAL NETWORKS
WPC2017 10
0.42 1 0.5 [0.25][<20.000][21.000-30.000][31.000-40][41.000-60.000]…
Age Gender
/= 100 /4
[0,100] 0 = Male, 0.25 = Female, 0.5=Other, 0.75 = Prefer not to Say, 1 =
Unk
Age Gender Income City Political party
30 Male 38,000 New York Democrat
39 Female 42,000 Page Republican
24 Other 39,000 San Francisco Republican
51 Prefer not to say 71,000 Seattle Democrat
ADVANCED MODELLING OF NEURAL NETWORKS
WPC2017 11
Method 1-of-(C-1) effects-coding: Standard deviation
𝜎 =
1
𝑁
𝑖=1
𝑁
𝑥𝑖 − 𝜇 2
𝜇 = 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑜𝑓 𝑎𝑙𝑙 𝑣𝑎𝑙𝑢𝑒𝑠
Age Gender Income City Political party
30 Male 38,000 New York Democrat
39 Female 42,000 Page Republican
ADVANCED MODELLING OF NEURAL NETWORKS
WPC2017 12
𝜎 =
1
𝑁
𝑖=1
𝑁
𝑥𝑖 − 𝜇 2
age = (30 + 36 + 52 + 42) / 4 =
40.0
𝜎 =
30 − 40 2 + 36 − 40 2 52 − 40 2 42 − 40 2
4
= 8,12
Age Gender Income City Political party
30 Male 38,000 New York Democrat
39 Female 42,000 Page Republican
ADVANCED MODELLING OF NEURAL NETWORKS
WPC2017 13
𝑉′
=
(𝑉 − 𝑚𝑒𝑎𝑛)
𝑠𝑡𝑑 𝑑𝑒𝑣
𝑉′
input will be used in place of the original input
Having the age average is 40.0, standard deviation is 8.12, and our current value is 30.0:
30.0 =
(30 − 40)
8.12
= −1.23
Age Gender Income City Political party
30 Male 38,000 New York Democrat
39 Female 42,000 Page Republican
ADVANCED MODELLING OF NEURAL NETWORKS
WPC2017 14
One of parameters: Italian cities (About 8000)
𝑀𝑖𝑙𝑎𝑛𝑜 𝑇𝑜𝑟𝑖𝑛𝑜 𝑅𝑜𝑚𝑎 … 𝐶𝑎𝑡𝑎𝑛𝑖𝑎
213 = 8192Binary compression:
With 13 nodes we can map 8192 values
(Having the same meaning/context)
City Value
Milano 0,0,0,0,0,0,0,0,0,0,0,0,0,0
Torino 0,0,0,0,1,1,0,0,0,0,0,1,0,0
Catania 0,1,0,0,1,0,0,0,0,1,0,1,1,0
ADVANCED MODELLING OF NEURAL NETWORKS
WPC2017 15
−1,23 1 3,4 [0.25][0,3]
The model has a mapping ratio of 1:1 between concepts and the number of neurons.
Only 5 parameters!
Can be managed without neural networks by an IF,THEN sequence in the code
Age Gender Income City Political party
30 Male 38,000 New York Democrat
39 Female 42,000 Page Republican
ADVANCED MODELLING OF NEURAL NETWORKS
WPC2017 16
Data must be manipulated and made understandable by the
machine, not for the humans!
Vulnerabilities
VULNERABILITIES
WPC2017 18
• Let’s imagine that we run an auction website like Ebay. On our website, we want to
prevent people from selling prohibited items .
• Enforcing these kinds of rules are hard if you have millions of users. We could hire
hundreds of people to review every auction listing by hand, but that would be
expensive.
VULNERABILITIES
WPC2017 19
Instead, we can use deep learning to automatically check auction photos for prohibited
items and flag the ones that violate the rules.
This is a typical image classification problem.
VULNERABILITIES – IMAGE CLASSIFICATION
WPC2017 20
We repeat this thousands of times with thousands of photos until the
model reliably produces the correct results with an acceptable accuracy.
VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS
WPC2017 21
• Convolutional neural networks are powerful models that consider the entire image
when classifying it.
• They can recognize complex shapes and patterns no matter where they appear in the
image.
• In many image recognition tasks, they can equal or even beat human performance.
VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS
WPC2017 22
With a fancy model like that, changing a few pixels
in the image to be darker or lighter shouldn’t have
a big effect on the final prediction, right?
Sure, it might change the final likelihood slightly,
but it shouldn’t flip an image from “prohibited” to
“allowed”.
“expectations”
VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS
WPC2017 23
It was discovered that this isn’t always true
VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS
WPC2017 24
If you know exactly which pixels to change and exactly how much to change them, you can
intentionally force the neural network to predict the wrong output for a given picture without
changing the appearance of the picture very much.
That means we can intentionally craft a picture that is clearly a prohibited item but which
completely fools our neural network
VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS
WPC2017 25
Why is this?
VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS
WPC2017 26
A machine learning classifier works by finding a dividing
line between the things it’s trying to tell apart.
Right now, the classifier works with 100% accuracy. It’s found
a line that perfectly separates all the green points from the
red points.
VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS
WPC2017 27
But what if we want to trick it into mis-classifying one of
the red points as a green point?
What’s the minimum amount we could move a red point
to push it into green territory?
If we add a small amount to the Y value of a red point
right beside the boundary, we can just barely push it
over into green territory.
VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS
WPC2017 28
In image classification with deep neural networks, each
“point” we are classifying is an entire image made up of
thousands of pixels.
That gives us thousands of possible values that we can
tweak to push the point over the decision line.
If we make sure that we tweak the pixels in the image in a
way that isn’t too obvious to a human, we can fool the
classifier without making the image look manipulated.
VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS
WPC2017 29
+ =
SquirelPeople
VULNERABILITIES – THE STEPS
WPC2017 30
1. Feed in the photo that we want to hack.
2. Check the neural network’s prediction and see
how far off the image is from the answer we
want to get for this photo.
3. Tweak our photo using back-propagation to
make the final prediction slightly closer to the
answer we want to get.
4. Repeat steps 1–3 a few thousand times with the
same photo until the network gives us the
answer we want.
VULNERABILITIES
Snippet of a Python script using Keras
HOW CAN WE PROTECT OURSELVES AGAINST
THESE ATTACKS?
WPC2017 32
Simply create lots of hacked images and include them in your training data set going
forward, that seems to make your neural network more resistant to these attacks.
This is called Adversarial Training and is probably the most reasonable defense to
consider adopting right now.
Thanks!

More Related Content

Similar to Deep learning italia speech galazzo

Machine learning and vulnerabilities
Machine learning and vulnerabilitiesMachine learning and vulnerabilities
Machine learning and vulnerabilities
galazzo
 
DSDT meetup July 2021
DSDT meetup July 2021DSDT meetup July 2021
DSDT meetup July 2021
DSDT_MTL
 
PyDresden 20170824 - Deep Learning for Computer Vision
PyDresden 20170824 - Deep Learning for Computer VisionPyDresden 20170824 - Deep Learning for Computer Vision
PyDresden 20170824 - Deep Learning for Computer Vision
Alex Conway
 
Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...
Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...
Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...
Alex Conway
 
V2 v posenet
V2 v posenetV2 v posenet
V2 v posenet
NAVER Engineering
 
[Revised] Intro to CNN
[Revised] Intro to CNN[Revised] Intro to CNN
[Revised] Intro to CNN
Vincent Tatan
 
Deep Learning for Computer Vision - ExecutiveML
Deep Learning for Computer Vision - ExecutiveMLDeep Learning for Computer Vision - ExecutiveML
Deep Learning for Computer Vision - ExecutiveML
Alex Conway
 
Transformer in Vision
Transformer in VisionTransformer in Vision
Transformer in Vision
Sangmin Woo
 
Neural Networks by Priyanka Kasture
Neural Networks by Priyanka KastureNeural Networks by Priyanka Kasture
Neural Networks by Priyanka Kasture
Priyanka Kasture
 
Lecture5:Social Network Analysis-By Dr. Y. Narahari
Lecture5:Social Network Analysis-By Dr. Y. NarahariLecture5:Social Network Analysis-By Dr. Y. Narahari
Lecture5:Social Network Analysis-By Dr. Y. Narahari
municsaa
 
Lecture5:Social Network Analysis-By Dr. Y. Narahari
Lecture5:Social Network Analysis-By Dr. Y. NarahariLecture5:Social Network Analysis-By Dr. Y. Narahari
Lecture5:Social Network Analysis-By Dr. Y. Narahari
municsaa
 
Review A DCNN APPROACH FOR REAL TIME UNCONSTRAINED FACE.pptx
Review A DCNN APPROACH FOR REAL TIME UNCONSTRAINED FACE.pptxReview A DCNN APPROACH FOR REAL TIME UNCONSTRAINED FACE.pptx
Review A DCNN APPROACH FOR REAL TIME UNCONSTRAINED FACE.pptx
AravindHari22
 
Neural Architectures for Still Images - Xavier Giro- UPC Barcelona 2019
Neural Architectures for Still Images - Xavier Giro- UPC Barcelona 2019Neural Architectures for Still Images - Xavier Giro- UPC Barcelona 2019
Neural Architectures for Still Images - Xavier Giro- UPC Barcelona 2019
Universitat Politècnica de Catalunya
 
Camp IT: Making the World More Efficient Using AI & Machine Learning
Camp IT: Making the World More Efficient Using AI & Machine LearningCamp IT: Making the World More Efficient Using AI & Machine Learning
Camp IT: Making the World More Efficient Using AI & Machine Learning
Krzysztof Kowalczyk
 
Human-Centered AI: Scalable, Interactive Tools for Interpretation and Attribu...
Human-Centered AI: Scalable, Interactive Tools for Interpretation and Attribu...Human-Centered AI: Scalable, Interactive Tools for Interpretation and Attribu...
Human-Centered AI: Scalable, Interactive Tools for Interpretation and Attribu...
polochau
 
BULK IEEE PROJECTS IN MATLAB ,BULK IEEE PROJECTS, IEEE 2015-16 MATLAB PROJEC...
 BULK IEEE PROJECTS IN MATLAB ,BULK IEEE PROJECTS, IEEE 2015-16 MATLAB PROJEC... BULK IEEE PROJECTS IN MATLAB ,BULK IEEE PROJECTS, IEEE 2015-16 MATLAB PROJEC...
BULK IEEE PROJECTS IN MATLAB ,BULK IEEE PROJECTS, IEEE 2015-16 MATLAB PROJEC...
Nexgen Technology
 
final year ieee pojects in pondicherry,bulk ieee projects ,bulk 2015-16 i...
  final  year ieee pojects in pondicherry,bulk ieee projects ,bulk  2015-16 i...  final  year ieee pojects in pondicherry,bulk ieee projects ,bulk  2015-16 i...
final year ieee pojects in pondicherry,bulk ieee projects ,bulk 2015-16 i...
nexgentech
 
深度學習在AOI的應用
深度學習在AOI的應用深度學習在AOI的應用
深度學習在AOI的應用
CHENHuiMei
 
Deep Residual Hashing Neural Network for Image Retrieval
Deep Residual Hashing Neural Network for Image RetrievalDeep Residual Hashing Neural Network for Image Retrieval
Deep Residual Hashing Neural Network for Image Retrieval
Edwin Efraín Jiménez Lepe
 
210523 swin transformer v1.5
210523 swin transformer v1.5210523 swin transformer v1.5
210523 swin transformer v1.5
taeseon ryu
 

Similar to Deep learning italia speech galazzo (20)

Machine learning and vulnerabilities
Machine learning and vulnerabilitiesMachine learning and vulnerabilities
Machine learning and vulnerabilities
 
DSDT meetup July 2021
DSDT meetup July 2021DSDT meetup July 2021
DSDT meetup July 2021
 
PyDresden 20170824 - Deep Learning for Computer Vision
PyDresden 20170824 - Deep Learning for Computer VisionPyDresden 20170824 - Deep Learning for Computer Vision
PyDresden 20170824 - Deep Learning for Computer Vision
 
Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...
Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...
Convolutional Neural Networks for Image Classification (Cape Town Deep Learni...
 
V2 v posenet
V2 v posenetV2 v posenet
V2 v posenet
 
[Revised] Intro to CNN
[Revised] Intro to CNN[Revised] Intro to CNN
[Revised] Intro to CNN
 
Deep Learning for Computer Vision - ExecutiveML
Deep Learning for Computer Vision - ExecutiveMLDeep Learning for Computer Vision - ExecutiveML
Deep Learning for Computer Vision - ExecutiveML
 
Transformer in Vision
Transformer in VisionTransformer in Vision
Transformer in Vision
 
Neural Networks by Priyanka Kasture
Neural Networks by Priyanka KastureNeural Networks by Priyanka Kasture
Neural Networks by Priyanka Kasture
 
Lecture5:Social Network Analysis-By Dr. Y. Narahari
Lecture5:Social Network Analysis-By Dr. Y. NarahariLecture5:Social Network Analysis-By Dr. Y. Narahari
Lecture5:Social Network Analysis-By Dr. Y. Narahari
 
Lecture5:Social Network Analysis-By Dr. Y. Narahari
Lecture5:Social Network Analysis-By Dr. Y. NarahariLecture5:Social Network Analysis-By Dr. Y. Narahari
Lecture5:Social Network Analysis-By Dr. Y. Narahari
 
Review A DCNN APPROACH FOR REAL TIME UNCONSTRAINED FACE.pptx
Review A DCNN APPROACH FOR REAL TIME UNCONSTRAINED FACE.pptxReview A DCNN APPROACH FOR REAL TIME UNCONSTRAINED FACE.pptx
Review A DCNN APPROACH FOR REAL TIME UNCONSTRAINED FACE.pptx
 
Neural Architectures for Still Images - Xavier Giro- UPC Barcelona 2019
Neural Architectures for Still Images - Xavier Giro- UPC Barcelona 2019Neural Architectures for Still Images - Xavier Giro- UPC Barcelona 2019
Neural Architectures for Still Images - Xavier Giro- UPC Barcelona 2019
 
Camp IT: Making the World More Efficient Using AI & Machine Learning
Camp IT: Making the World More Efficient Using AI & Machine LearningCamp IT: Making the World More Efficient Using AI & Machine Learning
Camp IT: Making the World More Efficient Using AI & Machine Learning
 
Human-Centered AI: Scalable, Interactive Tools for Interpretation and Attribu...
Human-Centered AI: Scalable, Interactive Tools for Interpretation and Attribu...Human-Centered AI: Scalable, Interactive Tools for Interpretation and Attribu...
Human-Centered AI: Scalable, Interactive Tools for Interpretation and Attribu...
 
BULK IEEE PROJECTS IN MATLAB ,BULK IEEE PROJECTS, IEEE 2015-16 MATLAB PROJEC...
 BULK IEEE PROJECTS IN MATLAB ,BULK IEEE PROJECTS, IEEE 2015-16 MATLAB PROJEC... BULK IEEE PROJECTS IN MATLAB ,BULK IEEE PROJECTS, IEEE 2015-16 MATLAB PROJEC...
BULK IEEE PROJECTS IN MATLAB ,BULK IEEE PROJECTS, IEEE 2015-16 MATLAB PROJEC...
 
final year ieee pojects in pondicherry,bulk ieee projects ,bulk 2015-16 i...
  final  year ieee pojects in pondicherry,bulk ieee projects ,bulk  2015-16 i...  final  year ieee pojects in pondicherry,bulk ieee projects ,bulk  2015-16 i...
final year ieee pojects in pondicherry,bulk ieee projects ,bulk 2015-16 i...
 
深度學習在AOI的應用
深度學習在AOI的應用深度學習在AOI的應用
深度學習在AOI的應用
 
Deep Residual Hashing Neural Network for Image Retrieval
Deep Residual Hashing Neural Network for Image RetrievalDeep Residual Hashing Neural Network for Image Retrieval
Deep Residual Hashing Neural Network for Image Retrieval
 
210523 swin transformer v1.5
210523 swin transformer v1.5210523 swin transformer v1.5
210523 swin transformer v1.5
 

More from Deep Learning Italia

Machine Learning driven Quantum Optimization for Marketing
Machine Learning driven Quantum Optimization for MarketingMachine Learning driven Quantum Optimization for Marketing
Machine Learning driven Quantum Optimization for Marketing
Deep Learning Italia
 
Modelli linguistici da Eliza a ChatGPT P roblemi , fraintendimenti e prospettive
Modelli linguistici da Eliza a ChatGPT P roblemi , fraintendimenti e prospettiveModelli linguistici da Eliza a ChatGPT P roblemi , fraintendimenti e prospettive
Modelli linguistici da Eliza a ChatGPT P roblemi , fraintendimenti e prospettive
Deep Learning Italia
 
Transformers In Vision From Zero to Hero (DLI).pptx
Transformers In Vision From Zero to Hero (DLI).pptxTransformers In Vision From Zero to Hero (DLI).pptx
Transformers In Vision From Zero to Hero (DLI).pptx
Deep Learning Italia
 
Meetup Luglio - Operations Research.pdf
Meetup Luglio - Operations Research.pdfMeetup Luglio - Operations Research.pdf
Meetup Luglio - Operations Research.pdf
Deep Learning Italia
 
Meetup Giugno - c-ResUNET.pdf
Meetup Giugno - c-ResUNET.pdfMeetup Giugno - c-ResUNET.pdf
Meetup Giugno - c-ResUNET.pdf
Deep Learning Italia
 
MEETUP Maggio - Team Automata
MEETUP Maggio - Team AutomataMEETUP Maggio - Team Automata
MEETUP Maggio - Team Automata
Deep Learning Italia
 
MEETUP APRILE - Ganomaly - Anomaly Detection.pdf
MEETUP APRILE - Ganomaly - Anomaly Detection.pdfMEETUP APRILE - Ganomaly - Anomaly Detection.pdf
MEETUP APRILE - Ganomaly - Anomaly Detection.pdf
Deep Learning Italia
 
2022_Meetup_Mazza-Marzo.pptx
2022_Meetup_Mazza-Marzo.pptx2022_Meetup_Mazza-Marzo.pptx
2022_Meetup_Mazza-Marzo.pptx
Deep Learning Italia
 
Machine Learning Security
Machine Learning SecurityMachine Learning Security
Machine Learning Security
Deep Learning Italia
 
The science of can and can t e la computazione quantistica
The science of can and can t e la computazione quantisticaThe science of can and can t e la computazione quantistica
The science of can and can t e la computazione quantistica
Deep Learning Italia
 
Dli meetup moccia
Dli meetup mocciaDli meetup moccia
Dli meetup moccia
Deep Learning Italia
 
Pi school-dli-presentation de nobili
Pi school-dli-presentation de nobiliPi school-dli-presentation de nobili
Pi school-dli-presentation de nobili
Deep Learning Italia
 
Machine Learning Explanations: LIME framework
Machine Learning Explanations: LIME framework Machine Learning Explanations: LIME framework
Machine Learning Explanations: LIME framework
Deep Learning Italia
 
Explanation methods for Artificial Intelligence Models
Explanation methods for Artificial Intelligence ModelsExplanation methods for Artificial Intelligence Models
Explanation methods for Artificial Intelligence Models
Deep Learning Italia
 
Use Cases Machine Learning for Healthcare
Use Cases Machine Learning for HealthcareUse Cases Machine Learning for Healthcare
Use Cases Machine Learning for Healthcare
Deep Learning Italia
 
NLG, Training, Inference & Evaluation
NLG, Training, Inference & Evaluation NLG, Training, Inference & Evaluation
NLG, Training, Inference & Evaluation
Deep Learning Italia
 
Transformer Seq2Sqe Models: Concepts, Trends & Limitations (DLI)
Transformer Seq2Sqe Models: Concepts, Trends & Limitations (DLI)Transformer Seq2Sqe Models: Concepts, Trends & Limitations (DLI)
Transformer Seq2Sqe Models: Concepts, Trends & Limitations (DLI)
Deep Learning Italia
 
Towards quantum machine learning calogero zarbo - meet up
Towards quantum machine learning  calogero zarbo - meet upTowards quantum machine learning  calogero zarbo - meet up
Towards quantum machine learning calogero zarbo - meet up
Deep Learning Italia
 
Macaluso antonio meetup dli 2020-12-15
Macaluso antonio  meetup dli 2020-12-15Macaluso antonio  meetup dli 2020-12-15
Macaluso antonio meetup dli 2020-12-15
Deep Learning Italia
 
Data privacy e anonymization in R
Data privacy e anonymization in RData privacy e anonymization in R
Data privacy e anonymization in R
Deep Learning Italia
 

More from Deep Learning Italia (20)

Machine Learning driven Quantum Optimization for Marketing
Machine Learning driven Quantum Optimization for MarketingMachine Learning driven Quantum Optimization for Marketing
Machine Learning driven Quantum Optimization for Marketing
 
Modelli linguistici da Eliza a ChatGPT P roblemi , fraintendimenti e prospettive
Modelli linguistici da Eliza a ChatGPT P roblemi , fraintendimenti e prospettiveModelli linguistici da Eliza a ChatGPT P roblemi , fraintendimenti e prospettive
Modelli linguistici da Eliza a ChatGPT P roblemi , fraintendimenti e prospettive
 
Transformers In Vision From Zero to Hero (DLI).pptx
Transformers In Vision From Zero to Hero (DLI).pptxTransformers In Vision From Zero to Hero (DLI).pptx
Transformers In Vision From Zero to Hero (DLI).pptx
 
Meetup Luglio - Operations Research.pdf
Meetup Luglio - Operations Research.pdfMeetup Luglio - Operations Research.pdf
Meetup Luglio - Operations Research.pdf
 
Meetup Giugno - c-ResUNET.pdf
Meetup Giugno - c-ResUNET.pdfMeetup Giugno - c-ResUNET.pdf
Meetup Giugno - c-ResUNET.pdf
 
MEETUP Maggio - Team Automata
MEETUP Maggio - Team AutomataMEETUP Maggio - Team Automata
MEETUP Maggio - Team Automata
 
MEETUP APRILE - Ganomaly - Anomaly Detection.pdf
MEETUP APRILE - Ganomaly - Anomaly Detection.pdfMEETUP APRILE - Ganomaly - Anomaly Detection.pdf
MEETUP APRILE - Ganomaly - Anomaly Detection.pdf
 
2022_Meetup_Mazza-Marzo.pptx
2022_Meetup_Mazza-Marzo.pptx2022_Meetup_Mazza-Marzo.pptx
2022_Meetup_Mazza-Marzo.pptx
 
Machine Learning Security
Machine Learning SecurityMachine Learning Security
Machine Learning Security
 
The science of can and can t e la computazione quantistica
The science of can and can t e la computazione quantisticaThe science of can and can t e la computazione quantistica
The science of can and can t e la computazione quantistica
 
Dli meetup moccia
Dli meetup mocciaDli meetup moccia
Dli meetup moccia
 
Pi school-dli-presentation de nobili
Pi school-dli-presentation de nobiliPi school-dli-presentation de nobili
Pi school-dli-presentation de nobili
 
Machine Learning Explanations: LIME framework
Machine Learning Explanations: LIME framework Machine Learning Explanations: LIME framework
Machine Learning Explanations: LIME framework
 
Explanation methods for Artificial Intelligence Models
Explanation methods for Artificial Intelligence ModelsExplanation methods for Artificial Intelligence Models
Explanation methods for Artificial Intelligence Models
 
Use Cases Machine Learning for Healthcare
Use Cases Machine Learning for HealthcareUse Cases Machine Learning for Healthcare
Use Cases Machine Learning for Healthcare
 
NLG, Training, Inference & Evaluation
NLG, Training, Inference & Evaluation NLG, Training, Inference & Evaluation
NLG, Training, Inference & Evaluation
 
Transformer Seq2Sqe Models: Concepts, Trends & Limitations (DLI)
Transformer Seq2Sqe Models: Concepts, Trends & Limitations (DLI)Transformer Seq2Sqe Models: Concepts, Trends & Limitations (DLI)
Transformer Seq2Sqe Models: Concepts, Trends & Limitations (DLI)
 
Towards quantum machine learning calogero zarbo - meet up
Towards quantum machine learning  calogero zarbo - meet upTowards quantum machine learning  calogero zarbo - meet up
Towards quantum machine learning calogero zarbo - meet up
 
Macaluso antonio meetup dli 2020-12-15
Macaluso antonio  meetup dli 2020-12-15Macaluso antonio  meetup dli 2020-12-15
Macaluso antonio meetup dli 2020-12-15
 
Data privacy e anonymization in R
Data privacy e anonymization in RData privacy e anonymization in R
Data privacy e anonymization in R
 

Recently uploaded

06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
Timothy Spann
 
Palo Alto Cortex XDR presentation .......
Palo Alto Cortex XDR presentation .......Palo Alto Cortex XDR presentation .......
Palo Alto Cortex XDR presentation .......
Sachin Paul
 
一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理
一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理
一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理
nuttdpt
 
Population Growth in Bataan: The effects of population growth around rural pl...
Population Growth in Bataan: The effects of population growth around rural pl...Population Growth in Bataan: The effects of population growth around rural pl...
Population Growth in Bataan: The effects of population growth around rural pl...
Bill641377
 
My burning issue is homelessness K.C.M.O.
My burning issue is homelessness K.C.M.O.My burning issue is homelessness K.C.M.O.
My burning issue is homelessness K.C.M.O.
rwarrenll
 
Experts live - Improving user adoption with AI
Experts live - Improving user adoption with AIExperts live - Improving user adoption with AI
Experts live - Improving user adoption with AI
jitskeb
 
一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理
一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理
一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理
g4dpvqap0
 
A presentation that explain the Power BI Licensing
A presentation that explain the Power BI LicensingA presentation that explain the Power BI Licensing
A presentation that explain the Power BI Licensing
AlessioFois2
 
一比一原版(Chester毕业证书)切斯特大学毕业证如何办理
一比一原版(Chester毕业证书)切斯特大学毕业证如何办理一比一原版(Chester毕业证书)切斯特大学毕业证如何办理
一比一原版(Chester毕业证书)切斯特大学毕业证如何办理
74nqk8xf
 
STATATHON: Unleashing the Power of Statistics in a 48-Hour Knowledge Extravag...
STATATHON: Unleashing the Power of Statistics in a 48-Hour Knowledge Extravag...STATATHON: Unleashing the Power of Statistics in a 48-Hour Knowledge Extravag...
STATATHON: Unleashing the Power of Statistics in a 48-Hour Knowledge Extravag...
sameer shah
 
一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理
一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理
一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理
bopyb
 
一比一原版(Harvard毕业证书)哈佛大学毕业证如何办理
一比一原版(Harvard毕业证书)哈佛大学毕业证如何办理一比一原版(Harvard毕业证书)哈佛大学毕业证如何办理
一比一原版(Harvard毕业证书)哈佛大学毕业证如何办理
zsjl4mimo
 
ViewShift: Hassle-free Dynamic Policy Enforcement for Every Data Lake
ViewShift: Hassle-free Dynamic Policy Enforcement for Every Data LakeViewShift: Hassle-free Dynamic Policy Enforcement for Every Data Lake
ViewShift: Hassle-free Dynamic Policy Enforcement for Every Data Lake
Walaa Eldin Moustafa
 
一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理
一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理
一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理
nuttdpt
 
一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理
一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理
一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理
nyfuhyz
 
DSSML24_tspann_CodelessGenerativeAIPipelines
DSSML24_tspann_CodelessGenerativeAIPipelinesDSSML24_tspann_CodelessGenerativeAIPipelines
DSSML24_tspann_CodelessGenerativeAIPipelines
Timothy Spann
 
Learn SQL from basic queries to Advance queries
Learn SQL from basic queries to Advance queriesLearn SQL from basic queries to Advance queries
Learn SQL from basic queries to Advance queries
manishkhaire30
 
University of New South Wales degree offer diploma Transcript
University of New South Wales degree offer diploma TranscriptUniversity of New South Wales degree offer diploma Transcript
University of New South Wales degree offer diploma Transcript
soxrziqu
 
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
apvysm8
 
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
Timothy Spann
 

Recently uploaded (20)

06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
 
Palo Alto Cortex XDR presentation .......
Palo Alto Cortex XDR presentation .......Palo Alto Cortex XDR presentation .......
Palo Alto Cortex XDR presentation .......
 
一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理
一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理
一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理
 
Population Growth in Bataan: The effects of population growth around rural pl...
Population Growth in Bataan: The effects of population growth around rural pl...Population Growth in Bataan: The effects of population growth around rural pl...
Population Growth in Bataan: The effects of population growth around rural pl...
 
My burning issue is homelessness K.C.M.O.
My burning issue is homelessness K.C.M.O.My burning issue is homelessness K.C.M.O.
My burning issue is homelessness K.C.M.O.
 
Experts live - Improving user adoption with AI
Experts live - Improving user adoption with AIExperts live - Improving user adoption with AI
Experts live - Improving user adoption with AI
 
一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理
一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理
一比一原版(Glasgow毕业证书)格拉斯哥大学毕业证如何办理
 
A presentation that explain the Power BI Licensing
A presentation that explain the Power BI LicensingA presentation that explain the Power BI Licensing
A presentation that explain the Power BI Licensing
 
一比一原版(Chester毕业证书)切斯特大学毕业证如何办理
一比一原版(Chester毕业证书)切斯特大学毕业证如何办理一比一原版(Chester毕业证书)切斯特大学毕业证如何办理
一比一原版(Chester毕业证书)切斯特大学毕业证如何办理
 
STATATHON: Unleashing the Power of Statistics in a 48-Hour Knowledge Extravag...
STATATHON: Unleashing the Power of Statistics in a 48-Hour Knowledge Extravag...STATATHON: Unleashing the Power of Statistics in a 48-Hour Knowledge Extravag...
STATATHON: Unleashing the Power of Statistics in a 48-Hour Knowledge Extravag...
 
一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理
一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理
一比一原版(GWU,GW文凭证书)乔治·华盛顿大学毕业证如何办理
 
一比一原版(Harvard毕业证书)哈佛大学毕业证如何办理
一比一原版(Harvard毕业证书)哈佛大学毕业证如何办理一比一原版(Harvard毕业证书)哈佛大学毕业证如何办理
一比一原版(Harvard毕业证书)哈佛大学毕业证如何办理
 
ViewShift: Hassle-free Dynamic Policy Enforcement for Every Data Lake
ViewShift: Hassle-free Dynamic Policy Enforcement for Every Data LakeViewShift: Hassle-free Dynamic Policy Enforcement for Every Data Lake
ViewShift: Hassle-free Dynamic Policy Enforcement for Every Data Lake
 
一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理
一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理
一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理
 
一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理
一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理
一比一原版(UMN文凭证书)明尼苏达大学毕业证如何办理
 
DSSML24_tspann_CodelessGenerativeAIPipelines
DSSML24_tspann_CodelessGenerativeAIPipelinesDSSML24_tspann_CodelessGenerativeAIPipelines
DSSML24_tspann_CodelessGenerativeAIPipelines
 
Learn SQL from basic queries to Advance queries
Learn SQL from basic queries to Advance queriesLearn SQL from basic queries to Advance queries
Learn SQL from basic queries to Advance queries
 
University of New South Wales degree offer diploma Transcript
University of New South Wales degree offer diploma TranscriptUniversity of New South Wales degree offer diploma Transcript
University of New South Wales degree offer diploma Transcript
 
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
 
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Dat...
 

Deep learning italia speech galazzo

  • 1. Machine Learning Best practices and Vulnerabilities MACHINE LEARNING BEST PRACTICES AND VULNERABILITIES
  • 4. THE PERCEPTRON WPC2017 4 In machine learning, the perceptron is a binary classifier or a function which can decide whether or not an input, represented by a vector of numbers, belongs to some specific class. 𝑓 𝑥 = χ ( ⟨ w , x ⟩ + b ) w is a vector having weights of real values, while operator ⟨ ⋅ , ⋅ ⟩ is the scalar product, b is the 'bias’, a constant not related to any input value and χ ( y ) is the output function
  • 5. MAIN EVOLUTIONS 1. Easy way, Logistic Regression, Support Vector Machine • Pro: Easy and fast use use • Cons: Low accuracy (compared to neural networks) 2. Hard way, Neural Networks • Pro: If get convergence you gain a very high accuracy (State of the art) • Cons: Very difficult to model, a lot of experience is required WPC2017 5
  • 6. EASY WAY • Pseudo equation • 𝑥 ∗∝ +𝑦 ∗ 𝛽 + 𝑐 ∗ 𝛿+. . +𝑧 ∗ 𝜔 = (0,1) • #logisticregression #svm WPC2017 6
  • 8. ADVANCED MODELLING OF NEURAL NETWORKS WPC2017 8 Use case, provide a customer's willingness to vote a political party Age Gender Income City Political party 30 Male 38,000 New York Democrat 39 Female 42,000 Page Republican 24 Other 39,000 San Francisco Republican 51 Prefer not to say 71,000 Seattle Democrat
  • 9. ADVANCED MODELLING OF NEURAL NETWORKS WPC2017 9 > 20 parameters 0,17 18,24 25,35 36,45 46,60 > 60 𝑚𝑎𝑙𝑒 𝑓𝑒𝑚𝑎𝑙𝑒 𝑢𝑟𝑏𝑎𝑛 𝑟𝑢𝑟𝑎𝑙 [𝑠𝑢𝑏𝑢𝑟𝑏𝑎𝑛]..[democrat][Republican] Age Gender Income City Political party 30 Male 38,000 New York Democrat 39 Female 42,000 Page Republican 24 Other 39,000 San Francisco Republican 51 Prefer not to say 71,000 Seattle Democrat
  • 10. ADVANCED MODELLING OF NEURAL NETWORKS WPC2017 10 0.42 1 0.5 [0.25][<20.000][21.000-30.000][31.000-40][41.000-60.000]… Age Gender /= 100 /4 [0,100] 0 = Male, 0.25 = Female, 0.5=Other, 0.75 = Prefer not to Say, 1 = Unk Age Gender Income City Political party 30 Male 38,000 New York Democrat 39 Female 42,000 Page Republican 24 Other 39,000 San Francisco Republican 51 Prefer not to say 71,000 Seattle Democrat
  • 11. ADVANCED MODELLING OF NEURAL NETWORKS WPC2017 11 Method 1-of-(C-1) effects-coding: Standard deviation 𝜎 = 1 𝑁 𝑖=1 𝑁 𝑥𝑖 − 𝜇 2 𝜇 = 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑜𝑓 𝑎𝑙𝑙 𝑣𝑎𝑙𝑢𝑒𝑠 Age Gender Income City Political party 30 Male 38,000 New York Democrat 39 Female 42,000 Page Republican
  • 12. ADVANCED MODELLING OF NEURAL NETWORKS WPC2017 12 𝜎 = 1 𝑁 𝑖=1 𝑁 𝑥𝑖 − 𝜇 2 age = (30 + 36 + 52 + 42) / 4 = 40.0 𝜎 = 30 − 40 2 + 36 − 40 2 52 − 40 2 42 − 40 2 4 = 8,12 Age Gender Income City Political party 30 Male 38,000 New York Democrat 39 Female 42,000 Page Republican
  • 13. ADVANCED MODELLING OF NEURAL NETWORKS WPC2017 13 𝑉′ = (𝑉 − 𝑚𝑒𝑎𝑛) 𝑠𝑡𝑑 𝑑𝑒𝑣 𝑉′ input will be used in place of the original input Having the age average is 40.0, standard deviation is 8.12, and our current value is 30.0: 30.0 = (30 − 40) 8.12 = −1.23 Age Gender Income City Political party 30 Male 38,000 New York Democrat 39 Female 42,000 Page Republican
  • 14. ADVANCED MODELLING OF NEURAL NETWORKS WPC2017 14 One of parameters: Italian cities (About 8000) 𝑀𝑖𝑙𝑎𝑛𝑜 𝑇𝑜𝑟𝑖𝑛𝑜 𝑅𝑜𝑚𝑎 … 𝐶𝑎𝑡𝑎𝑛𝑖𝑎 213 = 8192Binary compression: With 13 nodes we can map 8192 values (Having the same meaning/context) City Value Milano 0,0,0,0,0,0,0,0,0,0,0,0,0,0 Torino 0,0,0,0,1,1,0,0,0,0,0,1,0,0 Catania 0,1,0,0,1,0,0,0,0,1,0,1,1,0
  • 15. ADVANCED MODELLING OF NEURAL NETWORKS WPC2017 15 −1,23 1 3,4 [0.25][0,3] The model has a mapping ratio of 1:1 between concepts and the number of neurons. Only 5 parameters! Can be managed without neural networks by an IF,THEN sequence in the code Age Gender Income City Political party 30 Male 38,000 New York Democrat 39 Female 42,000 Page Republican
  • 16. ADVANCED MODELLING OF NEURAL NETWORKS WPC2017 16 Data must be manipulated and made understandable by the machine, not for the humans!
  • 18. VULNERABILITIES WPC2017 18 • Let’s imagine that we run an auction website like Ebay. On our website, we want to prevent people from selling prohibited items . • Enforcing these kinds of rules are hard if you have millions of users. We could hire hundreds of people to review every auction listing by hand, but that would be expensive.
  • 19. VULNERABILITIES WPC2017 19 Instead, we can use deep learning to automatically check auction photos for prohibited items and flag the ones that violate the rules. This is a typical image classification problem.
  • 20. VULNERABILITIES – IMAGE CLASSIFICATION WPC2017 20 We repeat this thousands of times with thousands of photos until the model reliably produces the correct results with an acceptable accuracy.
  • 21. VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS WPC2017 21 • Convolutional neural networks are powerful models that consider the entire image when classifying it. • They can recognize complex shapes and patterns no matter where they appear in the image. • In many image recognition tasks, they can equal or even beat human performance.
  • 22. VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS WPC2017 22 With a fancy model like that, changing a few pixels in the image to be darker or lighter shouldn’t have a big effect on the final prediction, right? Sure, it might change the final likelihood slightly, but it shouldn’t flip an image from “prohibited” to “allowed”. “expectations”
  • 23. VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS WPC2017 23 It was discovered that this isn’t always true
  • 24. VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS WPC2017 24 If you know exactly which pixels to change and exactly how much to change them, you can intentionally force the neural network to predict the wrong output for a given picture without changing the appearance of the picture very much. That means we can intentionally craft a picture that is clearly a prohibited item but which completely fools our neural network
  • 25. VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS WPC2017 25 Why is this?
  • 26. VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS WPC2017 26 A machine learning classifier works by finding a dividing line between the things it’s trying to tell apart. Right now, the classifier works with 100% accuracy. It’s found a line that perfectly separates all the green points from the red points.
  • 27. VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS WPC2017 27 But what if we want to trick it into mis-classifying one of the red points as a green point? What’s the minimum amount we could move a red point to push it into green territory? If we add a small amount to the Y value of a red point right beside the boundary, we can just barely push it over into green territory.
  • 28. VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS WPC2017 28 In image classification with deep neural networks, each “point” we are classifying is an entire image made up of thousands of pixels. That gives us thousands of possible values that we can tweak to push the point over the decision line. If we make sure that we tweak the pixels in the image in a way that isn’t too obvious to a human, we can fool the classifier without making the image look manipulated.
  • 29. VULNERABILITIES - CONVOLUTIONAL NEURAL NETWORKS WPC2017 29 + = SquirelPeople
  • 30. VULNERABILITIES – THE STEPS WPC2017 30 1. Feed in the photo that we want to hack. 2. Check the neural network’s prediction and see how far off the image is from the answer we want to get for this photo. 3. Tweak our photo using back-propagation to make the final prediction slightly closer to the answer we want to get. 4. Repeat steps 1–3 a few thousand times with the same photo until the network gives us the answer we want.
  • 31. VULNERABILITIES Snippet of a Python script using Keras
  • 32. HOW CAN WE PROTECT OURSELVES AGAINST THESE ATTACKS? WPC2017 32 Simply create lots of hacked images and include them in your training data set going forward, that seems to make your neural network more resistant to these attacks. This is called Adversarial Training and is probably the most reasonable defense to consider adopting right now.

Editor's Notes

  1. Here’s how that looks on a graph for a simple two-dimensional classifier that’s learned to separate green points (acceptable) from red points (prohibited)
  2. Here’s how that looks on a graph for a simple two-dimensional classifier that’s learned to separate green points (acceptable) from red points (prohibited)
  3. Here’s how that looks on a graph for a simple two-dimensional classifier that’s learned to separate green points (acceptable) from red points (prohibited)
  4. Here’s how that looks on a graph for a simple two-dimensional classifier that’s learned to separate green points (acceptable) from red points (prohibited)
  5. Here’s how that looks on a graph for a simple two-dimensional classifier that’s learned to separate green points (acceptable) from red points (prohibited)
  6. Here’s how that looks on a graph for a simple two-dimensional classifier that’s learned to separate green points (acceptable) from red points (prohibited)