Day of AI 2024: 3-5
Essential Question: How Can Computers/ Machines Use Data To Understand The World?
Structure of Learning Activity
Explore Machine Object Recognition
Vocabulary (data, features, labels)
Feature Identification
Sorting and labeling activity
Teachable Machine demo
Teachebale Machine activity
Explore the data to train an AI model
Understand AI's carbon footprint
Vocabulary
Features: Unique attributes that identify an object
Data: Information we collect and use to answer questions and make decisions
Label: Text added to data to help train a machine learning model
Carbon Footprint: amount of carbon dioxide (CO2) emissions associated with all the activities of a person or other entity (e.g., building, corporation, country, etc.).
Materials Needed
Student Materials
Feature Table (1 per pair/group)
Sorting Sheet (1 per pair/group)
Pencils (1 per pair/group)
Manipulative set for each pair/group - LEGO or other manipulative objects with at least 3 different features
Teacher device for accessing:
Image files for the Image Recognition activity
Some of the files are created for the K-12 lesson and shared with the 3-5 lesson.
California K-12 CS Standards
3-5.DA.9 Use data to highlight and/or propose relationships, predict outcomes, or communicate ideas.
In addition to the above standard, this activity relates to Practice 4. Developing and Using Abstractions of the K–12 Computer Science Framework as students identify common features in a dataset.
AI4K12 Big Ideas in AI Guidelines
Processing (Sensing vs Perception)
3-5.1-B-i Use a software tool such as a speech transcription or visual object recognition demo to demonstrate machine perception, and explain why this is perception rather than mere sensing.
Representation (Feature vectors)
K-2.2-A-iv Identify the features that make each object in a collection unique. and create a table of features to organize the objects.
Nature of Learning (Finding patterns in data)
K-2.3-A-ii Identify patterns in labeled data and determine the features that predict labels.
Nature of Learning (Training a Model)
3-5.3-A-iii Train a classification model using machine learning, and then examine the accuracy of the model on new inputs.
Nature of Learning (Constructing vs. using a reasoner)
3-5.3-A-iv Demonstrate how training data are labeled when using a machine learning tool.
Datasets (Feature sets)
3-5.3-C-i Create a labeled dataset with explicit features of several types and use a machine learning tool to train a classifier on this data.
Datasets (Large datasets)
3.5.3-C-ii Illustrate how training a classifier for a broad concept such as "dog" requires a large amount of data to capture the diversity of the domain.