Extracting Relevant Questions to an RDF Dataset Using Formal Concept Analysis
ML RELEVANT FORMAL AND NON-FORMAL …
30
1 ML RELEVANT FORMAL AND NON-FORMAL EDUCATION/TRAINING OFEERINGS MACHINE LEARNING FUNDAMENTALS Qualification / Accreditation / Certificate Certificate of Completion Institution / Provider Telecom Academy Country Romania Type of course Non-formal Mode of study (check all that apply, e.g. in case of blended learning) x Class-based learning ☐ Work place-based learning ☒ Online course (live teaching / online individual study) ☐ Other type of distance learning Duration 1-4 months to complete depending on the chosen format Scope of course Fundamentals Target audience information ICT professionals (above beginner level) Core Modules / Topics addressed − Machine Learning general concepts − Linear regression − Naive Bayes algorithms − Data analysis and elements of verbal language processing − Support Vector Machine − K Means Knowledge/skills/compete nces (to be) obtained upon the completion of training Knowledge Upon successful completion of the program, learners will know/be familiar with: - classification algorithms - regression algorithms - clustering techniques - data cleaning / preparation techniques
Transcript of ML RELEVANT FORMAL AND NON-FORMAL …
1
ML RELEVANT FORMAL AND NON-FORMAL EDUCATION/TRAINING OFEERINGS
MACHINE LEARNING FUNDAMENTALS
Qualification /
Accreditation / Certificate
Certificate of Completion
Institution / Provider Telecom Academy
Country Romania
Type of course Non-formal
Mode of study
(check all that apply,
e.g. in case of blended
learning)
x Class-based learning
☐ Work place-based learning
☒ Online course (live teaching / online individual study)
☐ Other type of distance learning
Duration 1-4 months to complete depending on the chosen format
Scope of course Fundamentals
Target audience
information
ICT professionals (above beginner level)
Core Modules / Topics
addressed
− Machine Learning general concepts
− Linear regression
− Naive Bayes algorithms
− Data analysis and elements of verbal language processing
− Support Vector Machine
− K Means
Knowledge/skills/compete
nces (to be) obtained upon
the completion of training
Knowledge
Upon successful completion of the program, learners will know/be
familiar with:
- classification algorithms
- regression algorithms
- clustering techniques
- data cleaning / preparation techniques
2
- techniques for extracting features from data (spoken language
processing)
- model evaluation techniques
Skills / Competences
Upon successful completion of the program, learners will be able to
know:
- how to create a virtual environment in Python
- how to view data
- how to implement different algorithms in Python
- how to draw a model in Python
- how to make pipelines (processing, training, validation and
saving the model for production)
- how to work with the most used libraries in the field of Machine
Learning (ex: numpy, nltk, scikit-learn, pandas, etc.)
Source https://www.telacad.ro/cursuri/machine-learning-fundamentals/
3
MACHINE LEARNING (I)
Qualification /
Accreditation / Certificate
Certificate of Completion
Institution / Provider Machine learning courses
Country Romania
Type of course Non-formal
Mode of study
(check all that apply,
e.g. in case of blended
learning)
☐ Class-based learning
☐ Work place-based learning
☐ Online course (e.g. MOOC)
x Other type of distance learning - webinar
Duration 2 months
Scope of course Upskilling
Target audience
information
− Professionals who want to improve their respective knowledge.
Core Modules / Topics
addressed
− ML project lifecycle
− Best practices / Tips and tricks
− Linear models
− Naive Bayes
− NLP (natural language processing)
− Tree models
− Neural Networks
− Multilayer perceptrons
Knowledge/skills/compete
nces (to be) obtained upon
the completion of training
Knowledge
Upon successful completion of the program, learners will know/be
familiar with:
ML project lifecycle
− Problem definition
− Dataset preparation
− Feature selection / preparation
− Feature engineering
− Feature importance
4
− Model debugging
− Error analysis
− Strategies to avoid over-fitting
Best practices / Tips and tricks
Linear models
− Linear regression
− Logistical regression
Naive Bayes
NLP (natural language processing)
Tree models
− DecisionTrees
− RandomForests
− GradientBoostedTrees
Neural Networks
− Multilayer perceptrons
Source https://www.curs-ml.com/event-info/machine-learning-i-5
5
MACHINE LEARNING – DATA SCIENCE
Qualification /
Accreditation / Certificate
E-certificate
Institution / Provider NobleProg
Country Romania
Type of course Non-formal
Mode of study
(check all that apply,
e.g. in case of blended
learning)
☐ Class-based learning
☐ Work place-based learning
☒ Online course (e.g. MOOC)
☐ Other type of distance learning
Duration 21 hours
Scope of course Upskilling
Target audience
information
− Graduates and professionals who want to gain knowledge in the
field of Machine Learning
− Prerequisite: Knowledge and awareness of Machine Learning
fundamentals
Core Modules / Topics
addressed
1. Machine Learning introduction
2. ML libraries and programming languages
3. Testing ML algorithms
4. Data preparation
5. Classification
6. Regression and numerical prediction
7. Unsupervised learning
8. Text mining
9. Recommendations engines and collaborative filtering
10. Association pattern mining
11. Outlier Analysis
12. Machine Learning case study
Knowledge/skills/compete
nces (to be) obtained upon
the completion of training
Knowledge
Upon successful completion of the program, learners will know/be
familiar with:
6
- Machine Learning introduction
o Types of Machine learning – supervised vs
unsupervised learning
o From Statistical learning to Machine learning
o The Data Mining workflow:
▪ Business understanding
▪ Data Understanding
▪ Data preparation
▪ Modelling
▪ Evaluation
▪ Deployment
o Machine learning algorithms
o Choosing appropriate algorithm to the problem
o Overfitting and bias-variance tradeoff in ML
- ML libraries and programming languages
o Why use a programming language
o Choosing between R and Python
o Python crash course
o Python resources
o Python Libraries for Machine learning
o Jupyter notebooks and interactive coding
- Testing ML algorithms
o Generalization and overfitting
o Avoiding overfitting
▪ Holdout method
▪ Cross-Validation
▪ Bootstrapping
o Evaluating numerical predictions
▪ Measures of accuracy: ME, MSE, RMSE, MAPE
▪ Parameter and prediction stability
o Evaluating classification algorithms
▪ Accuracy and its problems
7
▪ The confusion matrix
▪ Unbalanced classes problem
o Visualizing model performance
▪ Profit curve
▪ ROC curve
▪ Lift curve
o Model selection
o Model tuning – grid search strategies
o Examples in Python
- Data preparation
o Data import and storage
o Understand the data – basic explorations
o Data manipulations with pandas library
o Data transformations – Data wrangling
o Exploratory analysis
o Missing observations – detection and solutions
o Outliers – detection and strategies
o Standarization, normalization, binarization
o Qualitative data recoding
o Examples in Python
- Classification
o Binary vs multiclass classification
o Classification via mathematical functions
▪ Linear discriminant functions
▪ Quadratic discriminant functions
o Logistic regression and probability approach
o k-nearest neighbors
o Naïve Bayes
o Decision trees
▪ CART
▪ Bagging
▪ Random Forests
8
▪ Boosting
▪ Xgboost
o Support Vector Machines and kernels
▪ Maximal Margin Classifier
▪ Support Vector Machine
o Ensemble learning
o Examples in Python
- Regression and numerical prediction
o Least squares estimation
o Variables selection techniques
o Regularization and stability- L1, L2
o Nonlinearities and generalized least squares
o Polynomial regression
o Regression splines
o Regression trees
o Examples in Python
- Unsupervised learning
o Clustering
▪ Centroid-based clustering – k-means, k-
medoids, PAM, CLARA
▪ Hierarchical clustering – Diana, Agnes
▪ Model-based clustering - EM
▪ Self organising maps
▪ Clusters evaluation and assessment
o Dimensionality reduction
▪ Principal component analysis and factor
analysis
▪ Singular value decomposition
o Multidimensional Scaling
o Examples in Python
- Text mining
o Preprocessing data
9
o The bag-of-words model
o Stemming and lemmization
o Analyzing word frequencies
o Sentiment analysis
o Creating word clouds
o Examples in Python
- Recommendations engines and collaborative filtering
o Recommendation data
o User-based collaborative filtering
o Item-based collaborative filtering
o Examples in Python
- Association pattern mining
o Frequent itemsets algorithm
o Market basket analysis
o Examples in Python
- Outlier Analysis
o Extreme value analysis
o Distance-based outlier detection
o Density-based methods
o High-dimensional outlier detection
o Examples in Python
- Machine Learning case study
o Business problem understanding
o Data preprocessing
o Algorithm selection and tuning
o Evaluation of findings
- Deployment
Source https://www.nobleprog.ro/cc/ml_lbg?participants=4&how=public
10
MACHINE LEARNING CRASH COURSE USING PYTHON
Qualification /
Accreditation
Certificate of Completion
Institution / Provider Esolutions Academy
Country Romania
Type of course Non-formal
Mode of study
(check all that apply,
e.g. in case of blended
learning)
☐ Class-based learning
☐ Work place-based learning
☒ Online course (e.g. MOOC)
☐ Other type of distance learning
Duration 2 days
Scope of course Upskilling
Target audience
information
Level of knowledge - any level
Core Modules / Topics
addressed
INTRO TO MACHINE LEARNING7
Linear Regression: A Visual Understanding
Minimizing Loss with Gradient Descent: A simple, visual approach
First Steps with Gradient Descents: A coding approach
Train, Test & Validation: How to split your data and why
Feature Engineering: Trying to make your model smarter
Underfitting versus Overfitting: Simple model versus complex model
LOGISTIC REGRESSION7
A Visual Understanding
Binary Classification: Tuning the “decision threshold”
Evaluation metrics: Precison & Recall
Understand Prediction Bias: Possible Root Causes
Predictions: Regularization in high dimensional space
From Logistic Regression to Neural Nets: Build your intuition
11
Knowledge/skills/compete
nces (to be) obtained upon
the completion of training
Knowledge
Upon successful completion of the program, learners will know/be
familiar with:
- All the concepts below will be combined with programming
exercises based on real datasets and use cases which will enable
the learner to have a clear and gradual understanding of
Machine Learning Skills / Competences
Upon successful completion of the program, learners will be able to:
- write their own Machine Learning applications
Source https://academy.esolutions.ro/course/machine-learning-crash-
course-using-python/
12
ARTIFICIAL NATURAL NETWORKS, MACHINE LEARNING, DEEP THINKING
Qualification /
Accreditation / Certificate
Certificate of completion
Institution / Provider bittnet
Country Romania
Type of course Non-formal
Mode of study
(check all that apply,
e.g. in case of blended
learning)
☒ Class-based learning
☐ Work place-based learning
☒ Online course (e.g. MOOC)
☒ Other type of distance learning – blended learning
Duration 3 days
Scope of course Advanced
Target audience
information
Developers
Prerequisite - Phyton
Core Modules / Topics
addressed
− Machine Learning Introduction – ML and Deep Learning classic
algorithms
− Python and Jupyter Notebook
− Numpy Python Library
− Pandas Python Library for Data Analysis
− Matplotlib Python Library for Data Visualization
− Seaborn Python Library for Data Visualization
− Supervised Learning
− Supervised Learning algorithms
− Unsupervised Learning
− Neural Networks
− Introduction to Tensorflow and Keras API
− Image Classification with Tensorflow
Knowledge/skills/compete
nces (to be) obtained upon
the completion of training
Knowledge
Upon successful completion of the program, learners will know/be
familiar with:
13
− Types of Machine Learning
− Virtual Environments
− Anaconda/miniconda/conda
− Arrays
− Indexing
− Operations
− Pandas Python Library for Data Analysis
− Basic plotting
− Saving plots
− Seaborn Python Library for Data Visualization
− Plots
− Splitting the dataset (train/val/test)
− Bias Variance Tradeoff
− Linear Regression, Polynomial Regression
− Model Evaluation, Selecting the Best Model
− Bias-Variance trade-off
− Logistic Regression
− Naive Bayes
− K Nearest Neighbors (KNN)
− Decision Trees and Random Forests
− Clustering
− K Means Clustering
− Dimensionality Reduction
− Neural Networks
− Tensorflow and Keras API
− MNIST Dataset
− Data Augmentation
Skills / Competences
Upon successful completion of the program, learners will be able to:
− understand and write Python
14
− Create beautiful sketches
− Upload, process and display data with Python Pandas (sources
include CSV, Excel, HTML, etc.)
− Understand the key concepts of machine learning and deep learning
− Build an image classifier for handwritten numbers and fashionable
clothes
Source https://www.bittnet.ro/tehnologii/artificial-intelligence/artificial-natural-
networks-machine-learning-deep-thinking/
15
DATA SCIENTIST CRYSTAL PROGRAM
Qualification /
Accreditation / Certificate
Certificate of Completion
Institution / Provider Crystal System in collaboration with the Politechnic University of Bucharest – Energetics Department
Country Romania
Type of course Non-formal
Mode of study
(check all that apply,
e.g. in case of blended
learning)
☐ Class-based learning
☐ Work place-based learning
☒ Online course (e.g. MOOC)
☐ Other type of distance learning
Duration
Scope of course Upskilling
Target audience
information
− Students interested in programming
Core Modules / Topics
addressed
− Data Science in a business context
− Math and statistics fundamentals for data Science
− R Programing for Data Science
− Data Science with R
− Python programming for data science
− Data Science with Python
− Artificial Intelligence and Machine Learning
− Tableau For data Science
− Big Data Hadoop and Spark Developer
− Data Science Case Study Project, real databases in various fields Knowledge/skills/compete
nces (to be) obtained upon
the completion of training
Knowledge
Upon successful completion of the program, learners will know/be
familiar with:
− The essential statements, constructs and idioms of Python and how
to develop and share their code using Jupyter notebooks.
− The basics of numpy and pandas libraries for querying in-memory
tabular data.
16
Skills / Competences
Upon successful completion of the program, learners will be able to:
− visualize the outcomes of data analyses using matplotlib and
seaborn.
− Process data on large clusters using PySpark
− Setup and assess machine learning models with scikit-learn.
− Communicate their analyses using informative graphics from
matplotlib and seaborn.
− Configure, train and assess machine learning models with scikit-
learn.
Source https://www.haec.gr/en/python-101
17
INTRODUCTION TO MACHINE LEARNING
Qualification /
Accreditation / Certificate
Certificate of Completion
Institution / Provider NobleProg
Country Romania
Type of course Non-formal
Mode of study
(check all that apply,
e.g. in case of blended
learning)
☐ Class-based learning
☐ Work place-based learning
☒ Online course (e.g. MOOC)
☐ Other type of distance learning
Duration 7 hours
Scope of course provide a practical introduction to machine learning for participants
interested in applying methods to the workplace.
Target audience
information
− This training course is intended for people who would like to apply
basic Machine Learning techniques in Machine Learning in practical
applications.
− Data researchers and statisticians who have some familiarity with
machine learning and know how to program the R program.
Core Modules / Topics
addressed
− Naive Bayes
− Multinomial models
− Bayesian categorical data analysis
− Discriminant analysis
− Linear regression
− Logistic regression
− GLM
− EM Algorithm
− Mixed Models
− Additive Models
− Classification
− KNN
− Ridge regression
18
− Clustering
Knowledge/skills/compete
nces (to be) obtained upon
the completion of training
Knowledge
Upon successful completion of the program, learners will know/be
familiar with:
− practical aspects of data / model preparation, execution, post-hoc
analysis, and visualization.
− practical introduction to machine learning for participants
interested in applying methods to the workplace
Source https://www.nobleprog.ro/cc/mlintro
19
MACHINE LEARNING FOR ROBOTICS
Qualification /
Accreditation / Certificate
Certificate of completiom
Institution / Provider NobleProg
Country Romania
Type of course Non-formal
Mode of study
(check all that apply,
e.g. in case of blended
learning)
☐ Class-based learning
☐ Work place-based learning
☒ Online course (e.g. MOOC)
☐ Other type of distance learning
Duration 21 hours
Scope of course Upskilling
Target audience
information
Prerequisite: High School maths, basics knowledge of statistics
Core Modules / Topics
addressed
− Regression
− Probabilistic Graphical Models
− Boosting
− Kernel Methods
− Gaussian Processes
− Evaluation and Model Selection
− Sampling Methods
− Clustering
− CRFs
− Random Forests
− IVMs
Knowledge/skills/compete
nces (to be) obtained upon
the completion of training
Knowledge
Upon successful completion of the program, learners will know/be
familiar with:
− This course introduces learning methods in robotics.
20
− It is an overview of the main methods, motivations and ideas
existing in the context of model recognition.
Skills / Competences
Upon successful completion of the program, learners will be able to:
− perform simple exercises using open source (usually R) or any other
popular software.
Source https://www.nobleprog.ro/cc/mlrobot1
21
INFORMATION SYSTEMS & SERVICES (BIG DATA AND ANALYTICS)
Qualification /
Accreditation / Certificate
Master Degree
Institution / Provider University of Piraeus (Department of Digital Systems)
Country Greece
Type of course Non-formal
Mode of study
(check all that apply,
e.g. in case of blended
learning)
☒ Class-based learning
☐ Work place-based learning
☒ Online course (e.g. MOOC)
☐ Other type of distance learning
Duration 1.5 years minimum
Scope of course Upskilling
Target audience
information
Scientists and professionals in areas of Digital Systems and Services
Core Modules / Topics
addressed
− Big Data and Analytics Ι: Techniques and Tools
− Machine Learning: Methods and Algorithms
− Big Data and Analytics ΙΙ: Techniques and Tools
− Business Process Analytics
− Cloud Computing with Hadoop
− Data Mining and Predictive Analytics
− Practical Machine Learning
− Big Data Analytics Applications
− MSc Dissertation
Knowledge/skills/compete
nces (to be) obtained upon
the completion of training
Knowledge
Upon successful completion of the program, learners will know/be
familiar with:
− Big Data and Analytics
− Machine Learning
− Business Process Analytics
− Cloud Computing with Hadoop
− Data Mining and Predictive Analytics
22
−
Skills / Competences
Upon successful completion of the program, learners will be able to:
− Apply their skills in the areas of: Big Data and Analytics, Machine
Learning, Business Process Analytics, Cloud Computing with
Hadoop, and Data Mining and Predictive Analytics
Source https://mscdss.ds.unipi.gr/en/about/big-data-analytics/
https://mscdss.ds.unipi.gr/en/about/
23
MACHINE LEARNING USING SPARK
Qualification /
Accreditation / Certificate
Certificate of completion
Institution / Provider Esolutions Acsdemy
Country Romania
Type of course Non-formal
Mode of study
(check all that apply,
e.g. in case of blended
learning)
☐ Class-based learning
☐ Work place-based learning
☒ Online course (e.g. MOOC)
☐ Other type of distance learning
Duration 3 days
Scope of course Upskilling
Target audience
information
− Prerequisites: IDE (Eclipse, IDEA)
Core Modules / Topics
addressed
− Spark and Kafka in the Big-Data Context
− Introduction to Apache Spark
− Interacting with Spark
− Programming with RDDs
− Working with Key-Value Pairs
− Simple Applications using Spark
− Spark DataFrames/Datasets
− Spark SQL
− Machine learning basics
− Overview of Spark MLlib & ML
− Details about Spark MLlib
− Performing linear algebra
− Scaling and normalizing features
− Training and applying a linear regression model
− Evaluating the model’s performance
− Using regularization
24
− Optimizing linear regression
− Recommendation Engines
− Details about Spark ML
− Logistic regression
− Decision trees & Random forests
− K-means clustering
Knowledge/skills/compete
nces (to be) obtained upon
the completion of training
Skills / Competences
Upon successful completion of the program, learners will be able to:
Learn how to use the latest Apache Spark for machine learning on big
data
Source https://academy.esolutions.ro/course/machine-learning-using-spark/
25
MACHINE LEARNING FUNDAMENTALS
Qualification /
Accreditation / Certificate
Certificate of completion
Institution / Provider INSTITUTUL DE STUDII FINANCIARE
Country Romania
Type of course Non-formal
Mode of study
(check all that apply,
e.g. in case of blended
learning)
☐ Class-based learning
☐ Work place-based learning
☒ Online course (e.g. MOOC)
☐ Other type of distance learning
Duration 8 hours
Scope of course Upskilling
Target audience
information
− The course is addressed to people in the area of risk management /
actuarial, marketing, anti-fraud / security, process automation and
those who hold middle and top management positions in the area of
strategy and digital transformation.
Core Modules / Topics
addressed
− Introduction to machine learning and artificial intelligence
− Machine learning on the financial market
− Basic concepts in statistical theory
− Supervised learning: regression and classification
− Unsupervised learning: clustering and analysis of key components
− Machine learning: neural networks
− How to build a machine learning algorithm (ML)?
− Evaluation of an algorithm: complexity and errors
− Case study 01: ML in insurance and private pensions:
o personalized insurance
o fraud identification
− Case study 02: ML in capital markets:
o algorithm-based trading
o anticipating the evolution of the share price
− Case study 03: ML in business:
26
o personalized recommendations
o
o deep learning
− Feedback
Knowledge/skills/compete
nces (to be) obtained upon
the completion of training
− The “Machine Learning Fundamentals” course introduces statistical
and business concepts for understanding artificial intelligence
algorithms and how they can be applied in different areas of
financial markets.
− The main objective of the course is to provide students with an
understanding of artificial intelligence algorithms as well as how
they can apply in the business area. The course will have a
theoretical component, with emphasis on elements of statistical
theory and financial mathematics as well as a practical one in which
different use-cases will be analyzed / implemented.
Source https://platforma.isfin.ro/ro/group/1840
27
INTRODUCTION TO MACHINE LEARNING WITH PYTHON
Qualification /
Accreditation / Certificate
Certificate of completion
Institution / Provider ETA2U Training Center
Country Romania
Type of course Non-formal
Mode of study
(check all that apply,
e.g. in case of blended
learning)
☐ Class-based learning
☐ Work place-based learning
☒ Online course (e.g. MOOC)
☐ Other type of distance learning
Duration 20 hours
Scope of course Upskilling
Target audience
information
− This course is for beginners with ML and intermediate learning
experience in Python
Core Modules / Topics
addressed
− ML Framework:
o Problem framing
o Building and deploying a ML project
− ML Strategies:
o Data wrangling: collecting | cleaning | representing |
inspecting | transforming |
o Model training: reducing loss | overfitting | regularization
|evaluating performance |
− ML Algorithms:
o Classification
o Regression
o Clustering
Knowledge/skills/compete
nces (to be) obtained upon
the completion of training
Skills / Competences
Upon successful completion of the program, learners will be able to:
− learn the main strategies and algorithms of machine learning (ML)
and how to implement them in Python.
28
− understand how to efficiently prepare a ML project
Source https://www.eta2u.ro/index.php/training/30-cursuri-specialisti/1595-
introduction-to-machine-learning-with-python
29
MACHINE LEARNING
Qualification /
Accreditation / Certificate
Master Degree
Institution / Provider Politechnica University Timisoara
Country Romania
Type of course Formal
Mode of study
(check all that apply,
e.g. in case of blended
learning)
x Class-based learning
☐ Work place-based learning
☒ Online course
☐ Other type of distance learning
Duration 4 semesters
Scope of course Upskilling
Target audience
information
ICT professionals (above beginner level)
Core Modules / Topics
addressed
Core component
each semester, at least 2 courses must be elected out of the following:
Semester 1: Fundamentals of Machine Learning, Optimization in
Machine Learning, Deep Learning, Fundamentals of Computer Vision,
Big Data Analytics
Semester 2: Natural Language Processing, Probabilistic Graphical
Models, Reinforcement Learning, Computer Vision, Big Data
Visualization
Semester 3: Robotics, Text Mining and Recommender Systems,
Autonomous Driving, Advanced Computer Vision, Big Data in Health and
Bioinformatics
Elective component
each semester, 1 course must be elected out of the course list of any
Master program in the field of Computer and Information Technology
Research component
develops the research skills of the Master student over the four
semesters of the program. It
30
contains courses and activities directly related to research and Master
Thesis development
Knowledge/skills/compete
nces (to be) obtained upon
the completion of training
− Advanced knowledge of the main topics and problems in the field of
machine learning
− Knowledge of current technologies and abilities to select and apply
them in the development of machine learning projects
− Skills to critically analyze, innovate, and research
− Advanced abilities of communication, interdisciplinary cooperation,
and team project management
Source https://cs.upt.ro/sites/default/files/inline/docs/studii_master/Progra
m_Overview-Master_ML.pdf
https://cs.upt.ro/ro/education/master#mml_details
