2/3 days
1. Machine Learning Introduction – algoritmi clasici de ML si Deep Learning
Definitions
What is it used for
Machine Learning Pipeline
Types of Machine Learning
** Supervised Learning
** Unsupervised Learning
** Reinforcement Learning
Summary of tools that we need to work with (Python, Numpy, etc)
2. Python and Jupyter Notebook
Introduction
Python Crash course
Python Packages
Virtual Environments
Anaconda/miniconda/conda
Jupyter Notebook Introduction
3. Numpy Python Library
Introduction
Arrays
Indexing
Operations
4. Pandas Python Library for Data Analysis
Introduction
Panda Series
Panda Frames
Data Input
5. Matplotlib Python Library for Data Visualization
Introduction
Basic plotting
Saving plots
Loading, displaying images
6.Seaborn Python Library for Data Visualization
Introduction
Plots
** Distribution Plots
** Categorical Plots
** Matrix Plots
7.Supervised Learning
What is a dataset
Splitting the dataset (train/val/test)
** Notes on the ability to generalize (Generalization)
Feature selection
** K Means Clustering
Bias Variance Tradeoff
Overfitting
Underfitting
What is an outlier?
How do we perform? Confusion Matrix
8.Supervised Learning algorithms
sci-kit learn introduction
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
9.Unsupervised Learning
Clustering
K Means Clustering
Dimensionality Reduction
** Principal Components Analysis (PCA)
** Singular Value Decomposition (SVD)
10.Neural Networks
Definitions
** Neuron
** Multiple Neurons
** Multiple Layers
** Fully Connected Layers
** Other Types of layers
Common Tasks (Image Classification, Object Detection, Segmentation, etc)
Number of parameters
Common Architectures
11.Introduction to Tensorflow and Keras API
Tensors
Computation Graph
Visualizing the Graph
Training
12. Image Classification with Tensorflow
Building a simple architecture by hand
MNIST Dataset
Data Augmentation