
World No. 1 Data Science and AI in Healthcare Training Course
Online/ Virtual Training
1 - 4 March, 2021
AI Roadmap in Healthcare

Roadmap to Machine Learning
Led by:

Walid Semaan
Founder and President
Matrix TRC "Data Science and AI Academy
Course description:
This 4-day (3h per day) online live course is intendent to go through the complete roadmap that leads to the immense universe of Artificial Intelligence. It starts from ground zero, a quick overview on basic statistics and data analysis, pillars upon which both “supervised” and “unsupervised” Machine Learning evolved and became at their turn the inevitable backbone for all Artificial Intelligence algorithms behind successful AI applications.
In parallel to dozens of step-by-step case studies, the workshop will explore simultaneously, and under different angles, all Machine Learning algorithms applied on the famous generic must-know case study “Iris Flowers”, as well as on “Pulse”, a real healthcare data set related to hospitalized patients. This exclusive approach, will allow participants to compare all solutions and adapt the exact one that fits their need in real life.

Learning Outcome
By the end of this workshop, participants will be able to compare systematically all “predictive” and “exploratory” Machine Learning algorithms, and more importantly their efficient application in their daily businesses. It will also clarify the “reality” of Artificial Intelligence and the science behind it, such as Neural Networks, image processing, Natural Language Processing and how all of them relate in the design of an AI project.
Participants will also have a comparative processing of some case studies under a multitude of technologies, of which some are “open source” (R and Python) and “proprietary” ones (SAS - SPSS – STATISTICA – Excel), allowing them to have a professional opinion in comparing different technologies.
A complete exhaustive material will be delivered in the workshop, detailing all the necessary information for all the topics.

Key Topics
-
Roadmap to Machine Learning in brief
-
Data Vizualisation
-
Descriptive statistics
-
Data Anaysis
-
-
Unspervised Machine Learning
-
Principal Component Analysis
-
Clustering techniques
-
-
Natural Language Processing
-
Word2Vec
-
AI applications in the healthcare industry
-
Covid-19 and AI applications
-
Supervised Machine Learning
-
Multiple Regressions
-
Discriminant Analysis
-
Naive Bayes
-
Decision Trees
-
KNN
-
Support Vector Machines
-
-
Aritificial Neural Networks
-
The FFNN Network
-
Deep Learning models
-
Convolutional Neural Networks
-
Recurrent Neural Networks
-
LSTM
-



Who should attend?
-
Researchers
-
Statisticians
-
Data Analysts
-
Data Scientists
-
Machine Learning modelers
-
Deep Learning software developers
-
AI practitioners
-
Computer science and IT managers

Schedule:
Day 1
1stSession (1h30)
Part 1: The Roadmap to Artificial Intelligence
-
Statistics
-
Data Analysis
-
Machine Learning
-
Neural Networks
-
Deep Learning
-
NLP
Part 2: Prior knowledge before Machine Learning (Statistics and Data Analysis)
-
Data Vizualisation: Histograms, pie charts, heat maps, combo charts
-
Descriptive statistics: Mean, medan, mode, variance, standard deviaion
-
Data Analysis: t-test, F Anova, Chi Square and simple regressions
Part 3: Multiple Linear and Logistic Regressions(“Supervised“ ML)
-
Geometrical models
- Coefficients and model validation
- Coefficient of determination vs. correlation
-
Probabilistic prediction
- Coefficientsand model validation
- Odds ratio
2ndSession (1h30)
Part 1: Discriminant Analysis(“Supervised“ ML)
-
Discriminant vs. Classification functions
-
Mahalanobis Squared distances
-
Probabilistic models
Part 2: Decision Trees(“Supervised“ ML)
-
CART tree: growing, pruning and validating methods
-
CHAID tree
-
Random Forest
Part 3: K Nearest Neighbors(“Supervised“ ML)
-
Types of neighboring distances
-
Rules of classification
-
Voting method
Day 2
1stSession (1h30)
Part 1: Naive Bayes(“Supervised“ ML)
-
Bayes Theorem
-
Naïve assumption
-
CClassification rule
Part 2: Support Vector Machines(“Supervised“ML)
-
Predictive model
-
Estimation model
-
Kernel trick
Part 3: Artificial Neural Networks(“Supervised“ ML)
-
Neurons, Hidden layers
-
Perceptron , Synapsis, Weights, Scores
-
Activation functions: Sigmoid, TanH, ReLU
-
Classification with the SoftMax rule
-
Google Neural Network live simulator
2ndSession (1h30)
Part 1: Convolutional Neural Networks(Deep Learning)
-
AI revolution with image recognition and captioning
-
Convolution layers: Strides / Features / Filters
-
ConvNet / Pads
-
Pooling and Fully connected layers
Part 2: Recurrent Neural Networks(Deep Learning)
-
The four types of RNNs
-
Flow of data
-
Short-term memory
-
Vanishing and exploding gradient
Part 3: Long- and Short- Term Memory (Deep Learning)
-
Cell state
-
Forget phase
-
Update phase
-
Convey phase
Day 3
1stSession (1h30)
Part 1: All is possible with Big Data and IoT(Big Data and IoT)
-
What is IoT?
-
M2M and embedded systems
-
What is big data?
-
BD ecosystem vs. traditional IT
-
The 5 Vs of big data
2ndSession (1h30)
Part 1: Principal Component Analysis(“Unsupervised“ ML)
-
Observations vs. Variables analysis
-
Quality of data reduction:Eigenvalues
-
Low vs. High number of observations
Part 2: Clustering Analysis(“Unsupervised“ ML)
-
Clustering process
-
Parameters: distances and agglomerative rules
-
Clustering on top of a PCA?
-
Non Hierarchical clustering
-
K Mean clustering
Day 4
1stSession (1h30)
Part 1: Natural Language Processing (NLP)
-
Fields of application: Sentiment Analysis, Text categorization …
-
Tokenization, Stemming, Lemmatization, POS tags
-
Named Entity Recognition
Part 2: Word Embeddings(NLP)
-
Hot and Word Vectors
-
Word2Vec: CBOW vs. SKIP GRAM
2ndSession (1h30)
Part 1: AI solutions in the Pharma / Healthcare industry
-
Managing medical records
-
Clinical diagnosis with image recognition
-
Digital nurses / Digital consultation
-
Health monitoring / Smart hospitals
Part 2: Case study of “Covid-19 ... last of the pandemics ?“
-
Data design
-
Collecting data with IoT
-
Channeling data with big data technologies
-
Analyzing data with machine learning
-
Streaming outputs with automated applications
About the Trainer
Walid Semaan is the founder and president of Matrix TRC "Data Science and AI Academy (certified by SAS international and partners with Abu Dhabi Business School). He graduated in engineering from Ecole Supérieure d’Ingénieurs à Beyrouth and holds a degree in finance and marketing from the Ecole Supérieure de Commerce de Paris (ESCP) and an MBA from the University Paris-Dauphine-Sorbonne in Paris.
He is the creator and architect of the automated analytical Artificial Intelligence behind “Triple One Analytics,“ winner of the Best Innovative ICT Project at the 2011 Arab Golden Chip Award. Walid is the trainer of all the workshops he continuously updates in parallel with the training and consulting delivered worldwide, related to research methodologies, data visualization, data analysis, machine learning, deep learning, and NLPs, as well as complete forecasting, statistical quality control, and epidemiology programs.
Walid joined lately VOSTAD Edx as an expert trainer in Data Science & AI and is an expert certified trainer in the Middle East for SAS, PWC Academy, MEIRC / Training PLUS, and Formatech.
In parallel, he holds thousands of hours teaching master programs at Saint Joseph University, assisting Ph.D. students in their advanced analytics and machine learning programs, and training local and international companies, to name few from hundreds: Central Bank of Lebanon (Lebanon), PWC (Dubai), SCAD (Abu Dhabi), OXY Petroleum (Oman), Government Statistics (Ajman - UAE), IPSOS (Lebanon), Dallah Hospital (KSA), Smart Dubai (Dubai), DarkMatter (Abu Dhabi), Indevco Industries (Lebanon and Egypt), Sanofi (Amsterdam).