Designing an Artificial Intelligent Avatar Agent to Help With Unsupervised Human Learning Sequence

Everyone's mind works a little differently, and thank heavens for that, other we'd all be like a bunch of bacteria with a plan to replicate and only one way of doing things, or worse, we'd all live in a beehive of humanity or be members of the Borg, all one, not individuality, no diversity, and…

Everyone's mind works a little differently, and thank heavens for that, other we'd all be like a bunch of bacteria with a plan to replicate and only one way of doing things, or worse, we'd all live in a beehive of humanity or be members of the Borg, all one, not individuality, no diversity, and no uniqueness; yuk! Whether you like Howard Gardner's theory on multiple intelligences or not is irrelevant here, but I do like to pique your curiosity with a topic involving the future of education online – let's talk.

Not long ago, I read an interesting paper which discussed Learning Sequencing which was aimed at helping those who design curriculums and those who wish to theorize on such learning models a bit about how to find the most efficient process. Personally, as an individual, I despise overbearing structure wherever it is in learning, overregulation in business, or a homeowners association trying to tell me what I can and can not do. Now then, after downloading about 20 research papers, I noted similar terms used in artificial intelligence machine learning theory.

No, that was not what I had intended to study, however, one of many papers on AI machine learning caught my eye, this was an interesting paper titled; “Prototype-Driven Learning for Sequence Models,” by Aria Haghighi and Dan Klein. In the abstract's first few sentences it stated;

“We investigate prototype-driven learning for primarily unsupervised sequence modeling. Prior knowledge is specified declaratively, by providing a few canonical examples of each target annotation label. This sparse prototype information is then propagated across a corpus using distributive similarity features in a log-linear generative model. ”

The rest of the abstract dealt completely with programming theory, though reading these first few sentences seems to have a lot of dual uses. Let me explain. You see, above I stated the frustration with learning in a box, by someone else's standards, time line, dictates, and plan. That clearly seems like a one-size fits all way of learning and seems to follow the “No Child Allowed to Advance” challenges we have in our schools today. However, what if we programmed AI systems to help the student design his or her own curriculum?

Now then, I'd like you to read the book; “The First 20 Hours: How to Learn Anything … Fast,” by Josh Kaufman or watch the video online (YouTube) titled; “The first 20 hours – how to learn anything: Josh Kaufman at TEDxCSU.”

Let me ask some more questions, if I might. What if we had each student take a survey which determined their strong suits of how their mind is formatted to best think? There are many to think, look up WikiPedia; “List of Thought Process” and what if we developed a strategy to tap into the students preferred method of thinking first, then allowed an AI software program to deliver them visuals, assignments, and information to coincide. Next, we watched their fast uptake of information and tested their interest and speed of uptake in the subject matter.

The AI ​​system would be creating prototype learning sequences for each individual human and learning what works best as it goes, based on the individual's learning style, which would also change over time. In doing so the AI ​​system would get close to perfect in providing the most efficient personalized curriculum. Now that's what I am talking about, and thanks for listening.