Skip main navigation

New offer! Get 30% off one whole year of Unlimited learning. Subscribe for just £249.99 £174.99. New subscribers only T&Cs apply

Find out more

Machine Learning Basics: Part1

v
14.3
Hello everyone in this class, I’m going to introduce the machine learning basics.
21.4
Machine learning is a subset of artificial intelligence. Machine learning algorithms build a statistical model based on sample data, which is also called “training data“. The training process is also called “learning”. The model can then used to make predictions or classification, such as predicting real-state prices, classifying dog types, and detecting objects. There are two diagrams that show the difference between machine learning and classical programming. About classical programming, all the functions and rules are defined and implemented by human experts. On the other hand, machine learning can automatically learn the rules from training data and answers. The machine learning flow can be roughly divided into four stages. The first stage is data collection.
70.6
We need to collect data first, and usually also need to label them for machines to learn. The second stage is to analyze our problem and select the right model for it. Once the model is chosen, we then need to define a method to evaluate the performance of our model, that is to say, to know how good our model is. So we define a loss function to evaluate our model. The lower the loss, the better the performance. Finally, we need an algorithm to train the parameters in our model. The training process is called optimization in machine learning terms. Generally speaking,
109.4
there are three types of machine learning: supervised learning, unsupervised learning and reinforcement learning. Let’s talk about the supervised learning and unsupervised learning first. Here is a question for you. Guess what is the difference between supervised and unsupervised learning? The difference is that supervised learning has a teacher to label the data. In other words, the learning process is supervised by humans. On the contrary, unsupervised learning do not need human to label the data. Therefore, unsupervised learning has more potential because it can remove the burden of labeling. However, supervised learning is still far more accurate and still the mainstream of machine learning.
152.1
Based on if data type are discrete or continuous, we can further classify supervised learning into classification or regression, and classify unsupervised learning into clustering and dimensionality reduction. We’ll talk about this in details later. Another important type of machine learning is reinforcement learning, which is about how software agents learning take actions in an environment to achieve maximum total rewards. DeepMind reinvented reinforcement learning by combining it with deep neural networks and created AlphaGo that won the 5-game Go match against 18-time world champion Lee Sedol, and started the machine learning fever. Here is the scikit-learn algorithm cheat-sheet. Scikit-learn is the most popular machine learning library in Python. Let’s take a look of this map.
204.7
From the start, the first criterion that we should check is “Do we have enough data?”. If we have less than 50 samples, go back and collect more data. The next step is to check if we have categorical data, Which is called discrete data. If we have, then we’ll do classification when data are labeled, or do clustering of unsupervised learning. If the data are continuous, then we’ll do regression or dimensionality reduction.

In this video, Prof. Lai, Kuan-Ting will teach the basics of machine learning.

Machine learning is a subset of artificial intelligence. Machine learning algorithms build a statistical model based on sample data, which is also called “training data“. The training process is also called “learning”. The model can then used to make predictions or classification, such as predicting real-state prices, classifying dog types, and detecting objects.

There are two diagrams that show the difference between machine learning and classical programming. About classical programming, all the functions and rules are defined and implemented by human experts. On the other hand, machine learning can automatically learn the rules from training data and answers.

Then he will talk about the machine learning flow chart. The machine learning flow can be roughly divided into four stages. In the next part, he will tall the classification of Machine learning.

This article is from the free online

Applications of AI Technology

Created by
FutureLearn - Learning For Life

Reach your personal and professional goals

Unlock access to hundreds of expert online courses and degrees from top universities and educators to gain accredited qualifications and professional CV-building certificates.

Join over 18 million learners to launch, switch or build upon your career, all at your own pace, across a wide range of topic areas.

Start Learning now