Machine Learning

Machine learning (ML) is a part of artificial intelligence that allows computers to learn and improve without being explicitly programmed (Jordan & Mitchell, 2015). It involves creating systems that analyze data, identify patterns, and make decisions on their own.

Over the past two decades, ML has become widely used in areas like computer vision, speech recognition, and robotics. Developers often train systems using examples rather than coding every possible response. This approach has made ML important in industries such as healthcare, finance, logistics, and marketing, where it is used for tasks like fraud detection and medical diagnostics (LeCun et al., 2015).

ML has also impacted scientific fields by enabling new ways to analyze large datasets in biology, cosmology, and social sciences. However, it raises concerns about data privacy, bias, and ethics (Mittelstadt et al., 2016). As ML continues to evolve, it promises to transform technology and solve complex problems. Its ability to adapt and learn makes it a key part of modern artificial intelligence.

Examples of Machine Learning in Action

In Google's Quick Draw game, you (the user) play a game of Pictionary with an AI. The user is given a word and asked to draw it, while the AI tries to guess the word before time runs out. If the AI cannot guess correctly in time, the correct answer is revealed. That drawing is then added to a collection of examples for that word, helping the AI learn to recognize patterns across similar drawings. Over time, this allows the AI to improve its ability to identify the same word drawn by different people.

Quick, Draw!Can a neural network learn to recognize doodles? See how well it does with your drawings and help teach it, just by playing.

With Google's Teachable Machine, you can train your own AI to recognize images or audio by classifying them into categories based on the data you provide. These sets of data are called "classes." When you create a class, you assign it a name so the AI can associate that label with the corresponding images or audio. After creating at least two classes, you can train and test your AI to see how well it can recognize new inputs. The website also features a variety of example projects that showcase what others have built using the tool.

Teachable MachineTrain a computer to recognize your own images, sounds, & poses. A fast, easy way to create machine learning models for your sites, apps, and more – no expertise or coding required.

Reflect

Once you have given either Quick Draw or Teachable Machine a try, reflect on the following questions:

How is the AI "learning"?
How much data does it take to train AI using machine learning?
What do you think would happen if you gave inaccurate or incorrect data while teaching the AI?

◄ Back: How AI Works

Next: Large Language Models ►

Page updated

Google Sites

Report abuse