22.4.1. Objective

Students will build a basic AI model using user-friendly tools. This project helps students understand the training and testing process, concept of datasets, and how machines learn patterns.

In this project, students are tasked with creating a basic AI model. This involves using easy-to-use tools that do not require advanced programming skills. The focus here is to help students grasp essential concepts like how to train an AI model using data, how to test the model's accuracy, and the importance of datasets that provide the information the AI needs to learn. By the end of this process, students will have a practical understanding of how AI learns from data and makes decisions based on patterns it identifies.

Think of teaching a child to recognize fruits. If you show them several pictures of apples and say 'This is an apple', they will learn to identify apples over time. Similarly, in AI modeling, we show the model many examples of data (like images or sounds) so it can learn to recognize patterns.

Tools to Use

1. Teachable Machine https://teachablemachine.withgoogle.com/ A web-based tool by Google that allows anyone to train a model using images, sounds, or poses.
2. Machine Learning for Kids https://machinelearningforkids.co.uk/ Designed especially for students to create and train models using text, images, or numbers and use them in Scratch or Python.

Two specific tools are recommended for this project. The first is Google’s 'Teachable Machine', a web-based platform that enables users to train AI models easily using various data types including images, sounds, and poses. This makes it accessible for students without a technical background. The second tool, 'Machine Learning for Kids', is tailored for children and offers a more structured approach to machine learning, allowing students to create and train models using either Scratch or Python, thereby integrating coding practices with machine learning concepts.

Imagine you are using a cooking app that guides you through making a dish step-by-step. Just like that, these tools guide students through the process of creating an AI model, breaking down complex machine learning tasks into simple, manageable steps that anyone can follow.

1. Choose a type of data (Images/Text/Sound).
2. Collect sample data for different classes.
3. Train the model using the tool.
4. Test and refine the model.
5. Present the output and share your learning.

The project unfolds through a series of straightforward steps. First, students must choose what type of data they want to work with—this could be images, text, or sounds. Next, they gather samples for different classifications (like different types of images for a model that needs to classify pictures). Following that, they train the AI model using the data collected. After training, it is essential to test the model to see how well it performs in making predictions or classifications, which may lead to refining based on the model's performance. Lastly, students are encouraged to present what they learned through this process, promoting communication and reflection.

Consider building a puzzle: first, you need to choose which puzzle you want to complete (data type). Then, you gather the pieces (data collection). After this, you start putting the pieces together to see what works (training). If some pieces don’t fit, you might need to adjust how you're putting them together (refine), and at the end, you show off your completed puzzle to others (presentation).

Example Ideas • Image Classifier: Recognize happy and sad faces. • Sound Classifier: Differentiate claps, whistles, and snaps. • Text Classifier: Identify positive or negative feedback.

Students can engage in different examples of AI projects that highlight basic AI functionalities. For instance, an image classifier project could train a model to differentiate between happy and sad facial expressions, allowing it to categorize emotional responses based on images. Another project idea includes sound classification, teaching the AI to identify different sounds like claps, whistles, and snaps, which can demonstrate how audio recognition works. Additionally, a text classifier could analyze written feedback, helping it recognize whether comments are positive or negative. These examples illustrate the diverse applications of AI in handling different categories of data.

Imagine having a pet that learns to recognize your friends by their faces, responds to certain sounds like 'dinner time', or can tell if you're happy or sad by reading your text messages. These projects encapsulate that concept of teaching an AI to understand and categorize data just like our pet learns from its surroundings.

--