Intro

We just learned the basics of AI including why to use AI, what is AI, how AI has been developed, basic machine learning concepts, processes and algorithms, and AI applications. The preparation knowledge from this week's study cast the first cornerstone to learning and doing AI.

Lab

1. Get familiar with the Anaconda environment (Spyder, Python, Scikit-Learn, Matplotlib). Write code to plot Y = X^2*sin(X), in which X=[0,0.1,...,4].
Hints:
In the process, you will import packages, try basic algebraic operations with NumPy arrays, and plot using Matplotlib.
2. Problem Solving with Scikit-learn: classification (using the Iris dataset) and regression (using the Diabetes dataset) tasks.
Hints:
2a. Load datasets (check https://scikit-learn.org/stable/datasets/toy_dataset.htmlLinks to an external site. to see how to load Diabetes and Linnerud separately). Check the structure of the data in Spyder and think about why they can be used for the intended classification/regression task.
e.g.,
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
2b. Select a machine learning algorithm (corresponding to a command/function in Scikit-Learning) to perform the classification/regression task.
Hints:
Check https://scikit-learn.org/stable/supervised_learning.html Links to an external site.for a list of algorithms/commands/functions. You can start with Support Vector Machines (SVC or SVR in sklearn.svm)
2c. Find a way to check the results (e.g., plot, evaluate using a metric (e.g., score method coming with SVC when using SVC for classification), and *try predictions with new data)

 

 

Intro

This week, we learned the most useful software tools for doing AI, including the coding language (AI), data processing and visualization (NumPy, Pandas, and Matplotlib), basic machine learning (Scikit-learn), and deep learning (TensorFlow and Keras). This is not the only option but represents one of the most popular and useful toolsets in the state of the practice.

Lab

1. Get familiar with the Anaconda environment (Spyder, Python, Scikit-Learn, Matplotlib). Write code to plot Y = X^2*sin(X), in which X=[0,0.1,...,4].
Hints:
In the process, you will import packages, try basic algebraic operations with NumPy arrays, and plot using Matplotlib.
2. Problem Solving with Scikit-learn: classification (using the Iris dataset) and regression (using the Diabetes dataset) tasks.
Hints:
2a. Load datasets (check https://scikit-learn.org/stable/datasets/toy_dataset.htmlLinks to an external site. to see how to load Diabetes and Linnerud separately). Check the structure of the data in Spyder and think about why they can be used for the intended classification/regression task.
e.g.,
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
2b. Select a machine learning algorithm (corresponding to a command/function in Scikit-Learning) to perform the classification/regression task.
Hints:
Check https://scikit-learn.org/stable/supervised_learning.html Links to an external site.for a list of algorithms/commands/functions. You can start with Support Vector Machines (SVC or SVR in sklearn.svm)
2c. Find a way to check the results (e.g., plot, evaluate using a metric (e.g., score method coming with SVC when using SVC for classification), and *try predictions with new data)

 

 

Intro

How do you like linear models? Hope this week's learning has helped you make a connection with some of the curve fitting work that you did before. If this is the first time that you get into contact with learning models and their applications in regression (or curve fitting), then please try to utilize this great tool in your future work. It is simple but very useful!

Lab

1. Get familiar with the Anaconda environment (Spyder, Python, Scikit-Learn, Matplotlib). Write code to plot Y = X^2*sin(X), in which X=[0,0.1,...,4].
Hints:
In the process, you will import packages, try basic algebraic operations with NumPy arrays, and plot using Matplotlib.
2. Problem Solving with Scikit-learn: classification (using the Iris dataset) and regression (using the Diabetes dataset) tasks.
Hints:
2a. Load datasets (check https://scikit-learn.org/stable/datasets/toy_dataset.htmlLinks to an external site. to see how to load Diabetes and Linnerud separately). Check the structure of the data in Spyder and think about why they can be used for the intended classification/regression task.
e.g.,
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
2b. Select a machine learning algorithm (corresponding to a command/function in Scikit-Learning) to perform the classification/regression task.
Hints:
Check https://scikit-learn.org/stable/supervised_learning.html Links to an external site.for a list of algorithms/commands/functions. You can start with Support Vector Machines (SVC or SVR in sklearn.svm)
2c. Find a way to check the results (e.g., plot, evaluate using a metric (e.g., score method coming with SVC when using SVC for classification), and *try predictions with new data)

 

 

Intro

You should have seen the how decision tree works including how to create a tree (training) and how to use it (testing). More importantly, the employment of information theory in the development of decision tree theories explains the foundation, power, and variations of this type of algorithm. Decision trees have been developed into more powerful random forests and such evolvement is not over.

Lab

1. Get familiar with the Anaconda environment (Spyder, Python, Scikit-Learn, Matplotlib). Write code to plot Y = X^2*sin(X), in which X=[0,0.1,...,4].
Hints:
In the process, you will import packages, try basic algebraic operations with NumPy arrays, and plot using Matplotlib.
2. Problem Solving with Scikit-learn: classification (using the Iris dataset) and regression (using the Diabetes dataset) tasks.
Hints:
2a. Load datasets (check https://scikit-learn.org/stable/datasets/toy_dataset.htmlLinks to an external site. to see how to load Diabetes and Linnerud separately). Check the structure of the data in Spyder and think about why they can be used for the intended classification/regression task.
e.g.,
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
2b. Select a machine learning algorithm (corresponding to a command/function in Scikit-Learning) to perform the classification/regression task.
Hints:
Check https://scikit-learn.org/stable/supervised_learning.html Links to an external site.for a list of algorithms/commands/functions. You can start with Support Vector Machines (SVC or SVR in sklearn.svm)
2c. Find a way to check the results (e.g., plot, evaluate using a metric (e.g., score method coming with SVC when using SVC for classification), and *try predictions with new data)

 

 

Intro

Now you know how SVM works. The basic idea of SVM and the implementation of SVM via coupling with the optimization theory is what requires a lot of attention. Try to get more practice in the use of these algorithms for addressing real-world problems to deepen your understanding.

Lab

1. Get familiar with the Anaconda environment (Spyder, Python, Scikit-Learn, Matplotlib). Write code to plot Y = X^2*sin(X), in which X=[0,0.1,...,4].
Hints:
In the process, you will import packages, try basic algebraic operations with NumPy arrays, and plot using Matplotlib.
2. Problem Solving with Scikit-learn: classification (using the Iris dataset) and regression (using the Diabetes dataset) tasks.
Hints:
2a. Load datasets (check https://scikit-learn.org/stable/datasets/toy_dataset.htmlLinks to an external site. to see how to load Diabetes and Linnerud separately). Check the structure of the data in Spyder and think about why they can be used for the intended classification/regression task.
e.g.,
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
2b. Select a machine learning algorithm (corresponding to a command/function in Scikit-Learning) to perform the classification/regression task.
Hints:
Check https://scikit-learn.org/stable/supervised_learning.html Links to an external site.for a list of algorithms/commands/functions. You can start with Support Vector Machines (SVC or SVR in sklearn.svm)
2c. Find a way to check the results (e.g., plot, evaluate using a metric (e.g., score method coming with SVC when using SVC for classification), and *try predictions with new data)

 

 

Intro

You should have got enough knowledge to understand the working mechanisms of ANN and its elements from weights, activation, and neurons to layers. In particular, the forward and backward propagation needs to be clearly assessed. If needed, spend more time on the deductions, and when possible, spend a little time trying to code an ANN from scratch, which will be very beneficial if you want to perform in-depth work for deep learning.

Lab

1. Get familiar with the Anaconda environment (Spyder, Python, Scikit-Learn, Matplotlib). Write code to plot Y = X^2*sin(X), in which X=[0,0.1,...,4].
Hints:
In the process, you will import packages, try basic algebraic operations with NumPy arrays, and plot using Matplotlib.
2. Problem Solving with Scikit-learn: classification (using the Iris dataset) and regression (using the Diabetes dataset) tasks.
Hints:
2a. Load datasets (check https://scikit-learn.org/stable/datasets/toy_dataset.htmlLinks to an external site. to see how to load Diabetes and Linnerud separately). Check the structure of the data in Spyder and think about why they can be used for the intended classification/regression task.
e.g.,
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
2b. Select a machine learning algorithm (corresponding to a command/function in Scikit-Learning) to perform the classification/regression task.
Hints:
Check https://scikit-learn.org/stable/supervised_learning.html Links to an external site.for a list of algorithms/commands/functions. You can start with Support Vector Machines (SVC or SVR in sklearn.svm)
2c. Find a way to check the results (e.g., plot, evaluate using a metric (e.g., score method coming with SVC when using SVC for classification), and *try predictions with new data)

 

 

Intro

Hope you have enjoyed this contact with the state-of-the-art deep learning techniques. While this area is still growing fast, it is essential to understand the key breakthroughs that can enable us to go deeper and wider in NN and those new NN architectures and innovations that help us address a variety of complicated tasks.

Lab

1. Get familiar with the Anaconda environment (Spyder, Python, Scikit-Learn, Matplotlib). Write code to plot Y = X^2*sin(X), in which X=[0,0.1,...,4].
Hints:
In the process, you will import packages, try basic algebraic operations with NumPy arrays, and plot using Matplotlib.
2. Problem Solving with Scikit-learn: classification (using the Iris dataset) and regression (using the Diabetes dataset) tasks.
Hints:
2a. Load datasets (check https://scikit-learn.org/stable/datasets/toy_dataset.htmlLinks to an external site. to see how to load Diabetes and Linnerud separately). Check the structure of the data in Spyder and think about why they can be used for the intended classification/regression task.
e.g.,
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
2b. Select a machine learning algorithm (corresponding to a command/function in Scikit-Learning) to perform the classification/regression task.
Hints:
Check https://scikit-learn.org/stable/supervised_learning.html Links to an external site.for a list of algorithms/commands/functions. You can start with Support Vector Machines (SVC or SVR in sklearn.svm)
2c. Find a way to check the results (e.g., plot, evaluate using a metric (e.g., score method coming with SVC when using SVC for classification), and *try predictions with new data)

 

 

Intro

Hope you have enjoyed this contact with the state-of-the-art deep learning techniques. While this area is still growing fast, it is essential to understand the key breakthroughs that can enable us to go deeper and wider in NN and those new NN architectures and innovations that help us address a variety of complicated tasks.
Hints:
In the process, you will import packages, try basic algebraic operations with NumPy arrays, and plot using Matplotlib.
2. Problem Solving with Scikit-learn: classification (using the Iris dataset) and regression (using the Diabetes dataset) tasks.
Hints:
2a. Load datasets (check https://scikit-learn.org/stable/datasets/toy_dataset.htmlLinks to an external site. to see how to load Diabetes and Linnerud separately). Check the structure of the data in Spyder and think about why they can be used for the intended classification/regression task.
e.g.,
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
2b. Select a machine learning algorithm (corresponding to a command/function in Scikit-Learning) to perform the classification/regression task.
Hints:
Check https://scikit-learn.org/stable/supervised_learning.html Links to an external site.for a list of algorithms/commands/functions. You can start with Support Vector Machines (SVC or SVR in sklearn.svm)
2c. Find a way to check the results (e.g., plot, evaluate using a metric (e.g., score method coming with SVC when using SVC for classification), and *try predictions with new data)

 

 

Intro

The bagging, boosting, and stacking ensemble learning algorithms provide us with innovative ways to create great learners from relatively bad ones. You should be able to differentiate them well and know how to implement them with tools like Scikit-learn. More practice with real-world problems and a better understanding of the parameters for implementing these algorithms will help you gain more out of ensemble learning.

Lab

1. Get familiar with the Anaconda environment (Spyder, Python, Scikit-Learn, Matplotlib). Write code to plot Y = X^2*sin(X), in which X=[0,0.1,...,4].
Hints:
In the process, you will import packages, try basic algebraic operations with NumPy arrays, and plot using Matplotlib.
2. Problem Solving with Scikit-learn: classification (using the Iris dataset) and regression (using the Diabetes dataset) tasks.
Hints:
2a. Load datasets (check https://scikit-learn.org/stable/datasets/toy_dataset.htmlLinks to an external site. to see how to load Diabetes and Linnerud separately). Check the structure of the data in Spyder and think about why they can be used for the intended classification/regression task.
e.g.,
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
2b. Select a machine learning algorithm (corresponding to a command/function in Scikit-Learning) to perform the classification/regression task.
Hints:
Check https://scikit-learn.org/stable/supervised_learning.html Links to an external site.for a list of algorithms/commands/functions. You can start with Support Vector Machines (SVC or SVR in sklearn.svm)
2c. Find a way to check the results (e.g., plot, evaluate using a metric (e.g., score method coming with SVC when using SVC for classification), and *try predictions with new data)

 

 

Intro

The clustering topic may appear slightly different from other topics in this course. Different clustering algorithms are common in the evaluation metrics but are different in other ways. You have gained a good understanding of how different algorithms work. If needed, please spend more time applying these algorithms to different types of data, which may help you better understand the applicability of different algorithms.

Lab

1. Get familiar with the Anaconda environment (Spyder, Python, Scikit-Learn, Matplotlib). Write code to plot Y = X^2*sin(X), in which X=[0,0.1,...,4].
Hints:
In the process, you will import packages, try basic algebraic operations with NumPy arrays, and plot using Matplotlib.
2. Problem Solving with Scikit-learn: classification (using the Iris dataset) and regression (using the Diabetes dataset) tasks.
Hints:
2a. Load datasets (check https://scikit-learn.org/stable/datasets/toy_dataset.htmlLinks to an external site. to see how to load Diabetes and Linnerud separately). Check the structure of the data in Spyder and think about why they can be used for the intended classification/regression task.
e.g.,
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
2b. Select a machine learning algorithm (corresponding to a command/function in Scikit-Learning) to perform the classification/regression task.
Hints:
Check https://scikit-learn.org/stable/supervised_learning.html Links to an external site.for a list of algorithms/commands/functions. You can start with Support Vector Machines (SVC or SVR in sklearn.svm)
2c. Find a way to check the results (e.g., plot, evaluate using a metric (e.g., score method coming with SVC when using SVC for classification), and *try predictions with new data)

 

 

Intro

We focused on two dimension reduction algorithms this week: PCA and LDA. Try to think and summarize their commonalities to better understand how dimension reduction does and can do. Also, make sure you can clearly tell the differences between these two methods. This will cast a foundation for you to understand other dimension reduction methods.

Lab

1. Get familiar with the Anaconda environment (Spyder, Python, Scikit-Learn, Matplotlib). Write code to plot Y = X^2*sin(X), in which X=[0,0.1,...,4].
Hints:
In the process, you will import packages, try basic algebraic operations with NumPy arrays, and plot using Matplotlib.
2. Problem Solving with Scikit-learn: classification (using the Iris dataset) and regression (using the Diabetes dataset) tasks.
Hints:
2a. Load datasets (check https://scikit-learn.org/stable/datasets/toy_dataset.htmlLinks to an external site. to see how to load Diabetes and Linnerud separately). Check the structure of the data in Spyder and think about why they can be used for the intended classification/regression task.
e.g.,
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
2b. Select a machine learning algorithm (corresponding to a command/function in Scikit-Learning) to perform the classification/regression task.
Hints:
Check https://scikit-learn.org/stable/supervised_learning.html Links to an external site.for a list of algorithms/commands/functions. You can start with Support Vector Machines (SVC or SVR in sklearn.svm)
2c. Find a way to check the results (e.g., plot, evaluate using a metric (e.g., score method coming with SVC when using SVC for classification), and *try predictions with new data)

 

 

Intro

We learn Anomaly Detection algorithms represented by PCA and LDA.

Lab

1. Get familiar with the Anaconda environment (Spyder, Python, Scikit-Learn, Matplotlib). Write code to plot Y = X^2*sin(X), in which X=[0,0.1,...,4].
Hints:
In the process, you will import packages, try basic algebraic operations with NumPy arrays, and plot using Matplotlib.
2. Problem Solving with Scikit-learn: classification (using the Iris dataset) and regression (using the Diabetes dataset) tasks.
Hints:
2a. Load datasets (check https://scikit-learn.org/stable/datasets/toy_dataset.htmlLinks to an external site. to see how to load Diabetes and Linnerud separately). Check the structure of the data in Spyder and think about why they can be used for the intended classification/regression task.
e.g.,
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
2b. Select a machine learning algorithm (corresponding to a command/function in Scikit-Learning) to perform the classification/regression task.
Hints:
Check https://scikit-learn.org/stable/supervised_learning.html Links to an external site.for a list of algorithms/commands/functions. You can start with Support Vector Machines (SVC or SVR in sklearn.svm)
2c. Find a way to check the results (e.g., plot, evaluate using a metric (e.g., score method coming with SVC when using SVC for classification), and *try predictions with new data)

 

 

Intro

Associationr rule learning investigates the binary relationship between different attributes. Three major algorthms are introduced this week.

Lab

1. Get familiar with the Anaconda environment (Spyder, Python, Scikit-Learn, Matplotlib). Write code to plot Y = X^2*sin(X), in which X=[0,0.1,...,4].
Hints:
In the process, you will import packages, try basic algebraic operations with NumPy arrays, and plot using Matplotlib.
2. Problem Solving with Scikit-learn: classification (using the Iris dataset) and regression (using the Diabetes dataset) tasks.
Hints:
2a. Load datasets (check https://scikit-learn.org/stable/datasets/toy_dataset.htmlLinks to an external site. to see how to load Diabetes and Linnerud separately). Check the structure of the data in Spyder and think about why they can be used for the intended classification/regression task.
e.g.,
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
2b. Select a machine learning algorithm (corresponding to a command/function in Scikit-Learning) to perform the classification/regression task.
Hints:
Check https://scikit-learn.org/stable/supervised_learning.html Links to an external site.for a list of algorithms/commands/functions. You can start with Support Vector Machines (SVC or SVR in sklearn.svm)
2c. Find a way to check the results (e.g., plot, evaluate using a metric (e.g., score method coming with SVC when using SVC for classification), and *try predictions with new data)

 

 

Intro

This week, we started with the basics of reinforcement learning including the MDK process. Make sure you can use this framework to model a real-world problems. Also, the implementation of typical simple value-based reinforcement learning algorithms using your own code will be a necessity for adequately mastering this topic. The use of RL tools like OpenAI Gym will also be a must-know to practice RL.

Lab

1. Get familiar with the Anaconda environment (Spyder, Python, Scikit-Learn, Matplotlib). Write code to plot Y = X^2*sin(X), in which X=[0,0.1,...,4].
Hints:
In the process, you will import packages, try basic algebraic operations with NumPy arrays, and plot using Matplotlib.
2. Problem Solving with Scikit-learn: classification (using the Iris dataset) and regression (using the Diabetes dataset) tasks.
Hints:
2a. Load datasets (check https://scikit-learn.org/stable/datasets/toy_dataset.htmlLinks to an external site. to see how to load Diabetes and Linnerud separately). Check the structure of the data in Spyder and think about why they can be used for the intended classification/regression task.
e.g.,
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
2b. Select a machine learning algorithm (corresponding to a command/function in Scikit-Learning) to perform the classification/regression task.
Hints:
Check https://scikit-learn.org/stable/supervised_learning.html Links to an external site.for a list of algorithms/commands/functions. You can start with Support Vector Machines (SVC or SVR in sklearn.svm)
2c. Find a way to check the results (e.g., plot, evaluate using a metric (e.g., score method coming with SVC when using SVC for classification), and *try predictions with new data)

 

 

Intro

This week, we started with the basics of reinforcement learning including the MDK process. Make sure you can use this framework to model a real-world problems. Also, the implementation of typical simple value-based reinforcement learning algorithms using your own code will be a necessity for adequately mastering this topic. The use of RL tools like OpenAI Gym will also be a must-know to practice RL.

Lab

1. Get familiar with the Anaconda environment (Spyder, Python, Scikit-Learn, Matplotlib). Write code to plot Y = X^2*sin(X), in which X=[0,0.1,...,4].
Hints:
In the process, you will import packages, try basic algebraic operations with NumPy arrays, and plot using Matplotlib.
2. Problem Solving with Scikit-learn: classification (using the Iris dataset) and regression (using the Diabetes dataset) tasks.
Hints:
2a. Load datasets (check https://scikit-learn.org/stable/datasets/toy_dataset.htmlLinks to an external site. to see how to load Diabetes and Linnerud separately). Check the structure of the data in Spyder and think about why they can be used for the intended classification/regression task.
e.g.,
from sklearn.datasets import load_diabetes
X, y = load_diabetes(return_X_y=True)
2b. Select a machine learning algorithm (corresponding to a command/function in Scikit-Learning) to perform the classification/regression task.
Hints:
Check https://scikit-learn.org/stable/supervised_learning.html Links to an external site.for a list of algorithms/commands/functions. You can start with Support Vector Machines (SVC or SVR in sklearn.svm)
2c. Find a way to check the results (e.g., plot, evaluate using a metric (e.g., score method coming with SVC when using SVC for classification), and *try predictions with new data)