Course Description: 

Teaching and learning are, in varying degrees, already responsive, or adaptive, to different learner needs. Parents seek to select the “right” tutor for each child. Teachers shift instruction when they see confused faces or heads resting on desks. Students call friends or search the Internet for help when stuck. Technology, though, promises to make this adaptivity smarter, more immediate, and much deeper. This course investigates different approaches to adaptive and personalized learning organized around the different ways that learners vary--cognitively, metacognitively, and emotionally. Through a series of two-week explorations and exercises, completed individually or in small groups, you will apply adaptive learning principles across a mix of audiences and contexts (e.g., informal adult learning, K-12, parent/child, and higher education). For example, in a middle school math classroom of 30 students, how do you determine the right next challenge or support for each individual learner? How do you monitor and develop focus and persistence for one young child versus another? How do you know if a learner is confused or bored and what response will be successful? How do you know which group of learners needs what intervention when? The two-week investigation format will allow you to complete at least one iteration per challenge, making an initial pass at your thinking in week one, collecting feedback, and making revisions for week two. You’ll get to think about and play with data collection, algorithm development, and reporting in accessible and creative ways, regardless of your technology background.

Goals

Here are some of the things I hope you will accomplish in T513: 

• develop a framework for considering the many ways learners vary, which variables matter in a range of contexts, and how to adapt learning and learning environments in response to those variables;
• design (at least in theory) some simple adaptive learning algorithms (no technology expertise required);
• consider how to collect data to illuminate variables that impact learning;
• explore how to help learners self-adapt in the face of learning obstacles;
• experiment with data reporting and mechanisms for how teacher/parent and technology can make each other smarter.

The Plan

The weekly outline below will continue to evolve as all elements of this new course are finalized.

1/27 Class 1: A Learner Variable Framework

We introduce a structure for organizing the different ways learners vary. 

Reading:

• Shute, V. J. & Zapata-Rivera, D. (2008). Adaptive technologies.  In J. M. Spector, D. Merrill, J. van Merriënboer, & M. Driscoll (Eds.), Handbook of Research on Educational Communications and Technology (3rd Edition)(pp. 277-294). New York, NY: Lawrence Erlbaum Associates, Taylor & Francis Group. Although this piece is now 9 years old, it still provides a very helpful and accessible framework for adaptive learning.
• http://udlseries.udlcenter.org/presentations/learner_variability.html?plist=explore provides a look at learner variability from a UDL perspective.

Assignment for 2/3:

Create a persona for each of the targeted learners - a young child, a school group, and an individual adolescent/adult - we’ll be addressing. Highlight cognitive, metacognitive, affective, and other differences.

2/3 Class 2: How do we know?

We review and elaborate the personas. Then we begin identifying ways we can monitor the variation. How can we tell when a learner is bored, unskilled, lacking specific knowledge or understanding, distracted, unable to hear clearly, filled with self-doubt, and so on?

Reading:

• Measurement Matters: Assessing Personal Qualities Other Than Cognitive Ability for Educational Purposes, Angela L. Duckworth1 and David Scott Yeager. EDUCATIONAL RESEARCHERMay 2015  44 no. 4 237-251. 
• Learning to Notice - Noticing is a term often used in the context of adapting instruction on the fly in response to observed behaviors. This article comes from a math perspective, but you can find others related to language, art, and more.
• What can computers know? Here are two pieces that I think are interesting. A short piece from Valerie Shute describes stealth assessment: http://myweb.fsu.edu/vshute/pdf/sa.pdf. And here D'Mello and Calvo lay out a theoretical framework for affect detection: https://sites.google.com/site/sidneydmello/files/calvo-tac10.pdf?attredirects=0  

Assignments for 2/10:

• Update and revise your personas. We’re going to use them throughout the semester.
• Start identifying mechanisms for capturing variations that can impact learning. How and what can we measure, observe, and ask about?

2/10 Class 3: What do we do?

We have explored how learners vary and potential ways to identify (or notice) those variations. Now we investigate how to adapt to those variations in useful ways. We’ll start by looking at ways learning environments already adapt to us, how we adapt environment ourselves, and how we collaborate in trying to maximize learning opportunities. Expect an accessible guest presentation on how machines theoretically can learn about how we learn.

Reading:

• Social-Psychological Interventions in Education: They’re Not Magic -  This paper summarizes a number of studies related to modifying a learner’s affective state. You can explore other examples in the world of nudges at Harvard - http://cpl.hks.harvard.edu/big-faculty-research - or Stanford - https://sparq.stanford.edu/solutions.
• Formative Assessment. Dylan Wiliam’s work is as good a place as any to connect adaptive learning to formative assessment.
• It would be helpful to have some thoughts about ways to modify a wide range of potential learner behaviors. I reluctantly point you to Wikipedia for an overview - https://en.wikipedia.org/wiki/Behavior_modification. You can explore more from that base.

Guest Presentation: Alexander Rush, Assistant Professor, Harvard School of Engineering and Applied Sciences. How Machines Learn: The GO Example.

Assignment for 2/17: Continue collecting examples of different models of adaptivity across the framework in the table below.

2/17 & 2/24 Classes 4-5: Mini-Project 1 - Cognitive Variables

The next 8 weeks will be devoted to 4 mini-projects. Each 2-week mini-project will focus on applying our emerging sensibilities about adaptive learning to one of 4 categories of variation we’ve outlined, starting with cognitive variables. Here’s how it will work:

1. Class will divide into 3 groups representing the 3 different types of learners we began describing and collecting personas for in week 1 – a child, a group in school, and an individual adult or adolescent.
2. Working individually or in pairs (with another member in your group) you’ll describe 2 learning goals for your target learner(s). The learning goals must represent 2 of the following 3 goals – 1) mastering a skill or piece of knowledge; 2) understanding a concept by applying it to new contexts; 3) changing a behavior or disposition. For example, someone in the Child group might focus on counting objects to 10 (skill) and respecting the wishes of others (behavior). Someone in the School Group group could pick academic risk-taking (group behavior) and applying the concept of friction in multiple contexts (group understanding).
3. During the first week of the mini-project, you will make a first pass at identifying the cognitive learner variables that matter for each learning goal, how you can monitor those variables, and what action you might take in response.

4. During the second week of the mini-project, you’ll share your first iteration with the other members of your group. Revise and elaborate. Insights and examples from guest speakers and readings will further push your thinking.
5. For the next mini-project, your group will rotate to a different learner category. (If you want to stick with the same learner category, that's probably okay too.) You should experiment designing (on paper) at least one automated adaptive system for each mini-project.

Guest presentation: David Kuntz, Chief Research Officer, Knewton, Inc. 

Readings and References:

• https://www.knewton.com/ is a leading provider of adaptive learning systems. You can access a technical whitepaper here - http://learn.knewton.com/technical-white-paper.
• http://www.enlearn.org/ offers a different approach to adaptive learning.

3/3 & 3/10 Classes 6-7: Mini-Project 2 - Metacognitive Variables

Follow the same pattern as mini-project 1.

Guest presentations: Adam Seldow from Facebook on the Summit Learning Platform - https://www.summitlearning.org.

 

Readings and References:

• http://www.ibm.com/watson/education/. IBM’s application of Watson to education is very interesting. Check out the tools for collecting behavior in the classroom.
• https://www.summitlearning.org/. The personalized learning platform that Facebook created for Summit Schools has no adaptive intelligence. But it’s still personal. How does that work?

 

3/17 – NO CLASS. SPRING BREAK

 

3/24 & 3/31 Classes 8-9: Mini-Project 3 - Affective Variables

Follow the same pattern as mini-projects 1 and 2.

Guest presentation:Ryan Baker, Associate Professor, Penn GSE - http://scholar.gse.upenn.edu/baker. Ryan will share his work on teasing out learner affect (boredom, frustration, engagement, etc.) in MOOCS, while playing games, and more.

Readings and References:

• http://tessera.proexam.org/ is a relatively new assessment for tracking social emotional learning characteristics. Check out the mix of formats they use.
• https://www.transformingeducation.org/measuringmesh/. For comparison, look at the approach Transforming Education describes for measuring MESH (Mindsets, Essential Skills, and Habits).
• I'm repeating the article on affect detection: https://sites.google.com/site/sidneydmello/files/calvo-tac10.pdf?attredirects=0.
• And I'm adding a recent piece from Ryan Baker in preparation for his talk: http://www.columbia.edu/~rsb2162/bosch-pp-ijcai16-camera.pdf.

 

4/7 & 4/14 Classes 10-11: Mini-Project 4 - Other Variables

For this final mini-project, you can pick your learner category. Now we’re considering all the variables that didn’t quite fit into the other categories. These variables can be physical – visual or hearing impairment, limited mobility, severe cognitive or physical deficits – or environmental.

Readings and guest presentations TBD.

 

4/21 Last class. What matters and what works?

In this final class, we’ll reflect on what we’ve explored and learned. Each member of the class will present his or her “best of” the mini-projects. What are you most proud of? What most opened up your learning and thinking?