March 10, 2020: FLOE infrastructure Continuation

Present: Cindy, Colin, Justin, Jutta, Michelle, Ned, Philip, Ted,

Video Recording

Google Doc


Split into 3 topics

  1. Learners to discover their own learning approachesHow do we find the match?
    1. How are we going to recognize the match
  2. OERs
    1. Possibility to work with Gooru’s OER repository

      API Docs:

  3. Matching
    1. The feedback of whether matching results work or don’t work for learners will be reported to both learners and the system.

Next steps:

  1. Environmental scan: What learning resources are available. What attached to them
  2. Jutta will provide scenarios about what learners will ask for
  3. Jutta will provide some links for the prior work of the environmental scan

Jutta's input:


Here are some scenarios. These are not intended to be requirements that we should cover. They are intended to help us map out what we can and cannot do and how this fits into the FLOE infrastructure. One of the assertions of the project is that AI or smarts may not be appropriate or does more harm than good in many scenarios. What is a viable alternative to programs such as Squirrel AI? If the learners become the data scientists in their own learning, what are the accessibility barriers? Can we model more accessible tools? How far can we go in making it easier to find a learning experience that matches the learner’s changing understanding of what works best for them?


Learning goal: Learn how to hem my jeans.

Preferred method:  Watching people do it. (video demonstration)

I need captions for speech and audio in English. (English captions)

I need it broken down into steps so I can pause the demonstration and try out each step.


Learning goal: Introductory conversational Spanish.

Preferred method: 1. A game

  1. Includes opportunity to practice speaking Spanish
  2. A way to track and summarize what has been covered so far, and rewards to keep me motivated.

I need to control the interface using voice.


Learning goal: Read first chapter of Charlotte’s Web

Preferred method: On my reader.

Option to speak word, phrase or sentence aloud. Track the parts I couldn’t read.

With a glossary look-up of difficult words.

Clarifications regarding the Discovery and Exploration Layer:

-       The tool is for the student to discover and explore what works best for them

-       We are not optimizing. We are allowing opportunities to explore. This includes the opportunity to fail, make mistakes and play.

-       The preferences are not distinct from, but also go beyond the UIO or morphic preferences to encompass an undetermined set of preferences related to learning

-       The application of the preferences goes beyond the user interface and/or the browser rendering, to OER. This can mean topics covered, amenability to rendering according to the preferences, the existence of things like captioning and description for the content, etc..

-       The idea is to give the learner/student an opportunity to state a question, determine how they will answer the question, and refine or contextualize the answer.

Starting practical questions:

  1. How much manual metadata can we replace (see the list of accessibility metadata from the project with OER Commons, see: and and
  • Can we detect whether a resource has captioning, or description, is amenable to alternative layout and text styling, whether the images have alt-text etc. 
  • Can we detect the format of the resource, the language of the resource, the reading level, etc.

Other technical tasks related to this:

If learners are to become data scientists into their own learning, what tools are available to do this and how accessible are they? These include tools for:

  1. Linking up data sources including IOT sources
  2. “Visualizing” or manifesting data, creating data monitors and alerts (e.g., you have reached your target time on task)
  3. Analyzing results
  4. Mixing quantified data with qualitative data (stories from storytelling tool)
  5. Reporting findings