Lesson plan 1

Title: Computer Vision -Image classification and Supervised learning

Age: 5-6 yrs

Brief Description: In this lesson we cover the key AI concepts of computer vision, image classification and supervised learning.

In this activity, students are working with sourcing, sorting and classifying digital data and developing their understanding of how AI systems analyse digital data in the form of images. In this activity, students will be using the Cognimates visual programming software to test their AI vision model. In doing so, students are exposed to reading and following algorithms, as well as the opportunity to modify and create algorithms in Cognimates depending on their year level and ability.

Introduction seeing objects

Objectives:

• The students will learn how to label images into categories and how to train an AI algorithm.
• Students acquire, store and validate different types of data, and use a range of software to interpret and visualise data to create information
• Students could explore the use of vision apps (e.g. AI Poly) to detect objects in their environment.

Procedure: Training AI to see cats or dogs using Cognimates

In this part of the lesson, students will now learn how to train an AI algorithm to classify cats and dogs using images. Students will test their AI algorithm by creating a simple ‘guessing game’ using

Cognimates

Train a vision AI model

Requirements

1. Cognimates: a visual programming platform build as a Scratch extension by MIT Media Lab (creator – Stefania Druga). This is a free platform for everyone.
2. Clarifai vision API: Cognimates platform uses Clarifai vision API to build custom AI projects

Text instructions

1. Once you have an API key, go to Cognimates website, select Train modelsand then select Train visio
2. Fill out the information such as project name(e.g. cats vs dogs), API key, and the categories (e.g. cat, dog).
3. Upload or Drag & drop the images you have collected into each category as below (Note: minimum of 10 images from each category).
4. Select Train model to train your first AI model to recognise images.
5. Upload a new image or enter an image URL to test the AI model and select Predict. This will display the category recognised by the AI model and confidence of AI model to predict this output. If your algorithm does not correctly classify the image, then you will need to continue to train it with more data

Instructions: Develop an AI project using Cognimates

1. Select Code & play .in Cognimates home page Familiarise yourself with the Vision Training (green color) blocks that will be used to develop vision AI projects.

2.Use Set API key to and Choose image model blocks to link your Cognimates project with the vision AI model you built previously.

3.Use Search image using link block to identify the image if you have included an image from the web. Also, use What do you see in the photo? block to predict the category of the image

Resources: Online videos, lab, you tube

Curriculum link: https://youtu.be/f4QChm0UOJk

Assessment:

 Questionnaire

1.Why is this project is considered as a supervised learning approach?

2.How can we improve the AI model if it was unable to recognise new images?

3.What might biased images look like? How can we make sure we aren’t using biased data?

Powerpoint presentation in which students will reflect their learning giving examples from real lifefor ex(For example, if self-driving cars did not have the vision technology to recognise and differentiate pedestrians from other objects like traffic lights, other vehicles, buildings, road signs, they will not be able to drive on public roads.

Lesson plan 2

Title: NLP& AI- Text Classifiaction

Age: 5-6 yrs

Brief Description: Students will be generating, organising and sorting textual data and using tools. In doing so, students may be addressing some of the following content descriptors for data:

In this activity, students will be using the Cognimates visual programming software to test their AI text model. In doing so, students are exposed to reading and following algorithms, as well as the opportunity to modify and create algorithms in Cognimates depending on their year level and ability.

Objectives:

• Students to explore information systems used in society and to explain the purposes of technologies and how they are designed to meet personal, community or school needs
• Students acquire, store and validate different types of data, and use a range of software to interpret and visualise data to create information

• Students design, modify and follow simple algorithms involving sequences of steps, branching, and iteration

Activity :AI Travel assistant-A virtual trip advisor

Procedure:

The teacher poses questions to students as a whole class:

• If I was interested in seeing [Art, Dinosaurs, plants, animals] where should I go this weekend?
• I love being near water. Where are some places I could visit this weekend?
• I’d like to do something that doesn’t involve much talking this weekend. Where could I go?

The teacher picks one of the suggested locations and the class brainstorms together as many words or sentences that they can think of that are associated with that place and what they might do, see or hear. For example, for “Art Gallery”: paintings, statues, quiet, silent, walk, see art, artists, painters. The goal of this is to get students thinking about words that are associated with these places so that students can more easily participate in the plugged project. This is also a learning point for AI, as people will ask questions in many different ways, using many different words and so the more data and ideas we generate about a place of interest, the more likely that we are to build an AI that is able to make associations.

Students are given 3 different places to work with (in pairs or small groups). For example “Museum, Theme Park and Zoo”. They brainstorm as many words as they can think of associated for each location. Using these words, they then create as many different questions as they can think of that someone might ask if wanting to visit a place like that with those characteristics. For example, “Where is a quiet place to go?” or “I’d like to see paintings by famous artists”.

Students use their questions as the basis to train their AI Trip Advisor using the steps outlined below.

Building a virtual assistant

Students will train an AI algorithm/model to process text data on places, leisure activities. Students will test their AI model by creating a simple trip advisor project using Cognimates

Requirements

1. Cognimates: a visual programming platform build as a Scratch extension by MIT Media Lab (creator – Stefania Druga). This is a free platform for everyone.
2. uClassify API: Cognimates platform uses uClassify text classification API to build custom AI projects.

Text instructions

1. Once you have uClassify API keys, go to Cognimates website, select Train modelsand then select Train text.
2. Fill out the information such as project name (e.g. Trip advisor), Read & Write API keys, and the categories (e.g. Museum, Theme park, Zoo).
3. Type the questions you have collected in the classroom activity in each category as below (Note: minimum of 10 questions from each category).

4.Select Train model to train your first AI text model to recognise text.

5.Type a new sentence or question to test the AI model and select Predict. This will display the category recognised by the AI model and confidence of AI model to predict this output.

Instructions: Develop a text AI project using Cognimates

1. Select Code & play in Cognimates home page You need to familiarise yourself with Text Training (green color) blocks that will be used to develop text-based AI projects.
2. Use Set Read API key,Set Write API key, Choose text model and Set username blocks to link your Cognimates project with the text AI model you have built previously.
3. Use What kind of phrase is? block to predict the category (e.g. Museum) of the input text (e.g. How do I explore more about Australian caves?)

Resources: Online videos, lab, you tube

Curriculum links: https://youtu.be/_xcYi9JQpw

Assessment

Travel planning to create a project to recommend books from the library based on students