Demystifying the Machine Learning Revolution from the Technology to Your Classroom
Listen and learn : Snapshot
Sunday, June 23, 2:30–3:30 pm
Presentation 1 of 2
Beacon Technology Across the Curriculum
Dr. Scott Garrigan
Machine learning (ML) is changing our lives. It’s in phone, bank, Amazon and Netflix services--even in school software. This session demystifies how the five kinds of ML think and learn. Compare machine to human learning, you’ll learn five ways you can connect ML principles to help students learn deeply.
|Audience:||Curriculum/district specialists, Teachers, Principals/head teachers|
|Attendee devices:||Devices useful|
|Attendee device specification:||Laptop: Chromebook, Mac, PC
Tablet: Android, iOS, Windows
|Participant accounts, software and other materials:||The presentation website URL will provide access to all resources.|
|Focus:||Digital age teaching & learning|
|Topic:||Computer science and computational thinking|
|Subject area:||STEM/STEAM, Computer science|
|ISTE Standards:||For Educators:
While most of the western world has implemented AI and ML across industries and universities, K-12 has lagged behind. One reason is that there are few, if any, K-12 teachers and technologists who understand AI and ML. This session addresses this gap by educating key participants from K-12 institutions. This session has two purposes and two objectives:
Purpose 1. To demystify machine learning through simple explanations of the five different models of ML and how they work, think, and learn. The specific models are explained in the outline.
Objective 1. Participants will recognize the conceptually basic ways that the five machine learning models do what appears to be magic!
Purpose 2. To help teachers understand five ways they can use these powerful learning models to both inform themselves and their students about their changing world AND to apply that knowledge to classroom learning tasks.
Objective 2. Participants will be able to identify five ways to apply knowledge of ML to the K-12 classroom to prepare students for their future and to help them better learn normal curriculum concepts.
The presenter has done similar talks, webinars, and interviews for ISTE over the past year. He has developed and will have taught a graduate education course in K-12 importance of Artificial Intelligence and Machine Learning for Lehigh University’s College of Education. The ideas and activities from this course are used in this session.
Additional general background and critical need statement:
Everyone has heard that AI will change the world and that Machine Learning may soon rival human learning. Principals and superintendents are offered “AI-powered” software to help students learn. But no one explains to educators exactly what this means and how it works. It’s presented by the media and vendors as black-box magic. This session WILL demystify machine learning by showing how each of the five main machine learning algorithms work. The conceptual explanations will be such that every teacher and technologist can follow and understand from a high level. The underlying statistical mathematics and programming code will not even be mentioned.
This topic is critical to educator at this time for two reasons. First is that it is incumbent upon schools to help students understand the AI and ML technologies that will so dramatically affect their futures. Second is that the five underlying machine learning algorithms actually are similar to specific ways that humans think. Teachers can integrate learning activities in their classrooms that are powerful ways to learn as a human or a machine. Specific technology-empowered classroom activities will shared that relate to each of the five ML systems.
This conceptual understanding followed by hands-on integration into the classroom, machine learning will no longer be black-box magic for teachers or students. They will gain powerful mental tools to understand emerging technologies that will change their lives.
This is a fast-moving 30-minute session with compelling visuals that progressively unveil the mysteries of one kind of machine learning after another. The K-12 activities use common tools and accessible websites that participants can integrate into their own lessons and will be briefly but clearly demonstrated.
3 min - introduction of presenter, topic, and context
12 min – clear, brief, conceptual graphical explanation of each of the 5 Machine Learning systems used in the world today (will really be explained simply!):
- symbolism – inverse deduction using computational logic
- connectionism – backpropagation (neural network feedback)
- evolution – genetic programming & emergent behavior
- Bayesian – Bayesian inference and probabilities
- analogizers – support vector machine based on similar examples
3 min – connection of ML models to models of human learning and its appropriate place in the classroom and curriculum using the following generics (specifics will be shared for many subjects and levels)
10 min – specific K-12 activities that relate the kind of learning in each of the 5 ML models to classroom technology activities that integrate to many levels and subjects:
- symbolism -- logic models in math, science, programming (Scratch to Python)
- connectionism -- concept mapping CmapTools and continuous feedback
- evolution -- emergent & evolving phenomena (Breve, Microsoft genetic running man)
- Bayesian -- emphasis on probability & PREDICTIVE VALUE of theory (R, Mathematica)
- analogizers -- learning through story and analogy (story creation software)
2 min – Q & A and closure (questions are invited throughout the session)
The main source for the five ML models is Pedro Domingo’s landmark book “The Master Algorithm: How the quest for the ultimate learning machine will remake our world.” Domingo is a Computer Science professor who earned the highest award in data science, highly respected in his field.
The secondary source comprises a series of mini-MOOC courses from U.C. Berkely and Harvard on Machine Learning.
The tertiary source is from the presenter's new graduate course on "AI and Machine Learning for K-12 Educators."
The main source for the K-12 activities is the presenter’s implementation of the activities in his own K-12 and higher ed teaching. The presenter is a recognized expert in K-12 AL and Machine Learning.
Bringing Computer Science to Middle School Through Physical Computing and micro:bit
Improving Student Tech-Life Balance With Digital Wellness Month
Ten Big Data, AI, and Machine Learning Projects for the K12 Classroom