Regina Saphier: My dream: A “Self Driving” learning journey
Regina Saphier: My dream: A “Self Driving” learning journey
I love Coursera, even when I am focusing on other activities, like advising startups and watching Udacity’s lean startup course with Steve Blank, because it is the embodiment of something I have been waiting for. I started my Coursera blog last year specifically about my unique experiences and ideas regarding Coursera (it is nice to see that thousands find my independent blog from over 100 countries). I am primarily on Coursera to monitor the development, and as a bonus I am enjoying the free style learning (when I have the time). Today I am maintaining a wait and see position regarding Coursera and Udacity (the former group might want to think about the elimination of deadlines, and the latter group in my opinion should massively broaden their self-produced course range). It means that I am not taking any classes right now (I am auditing). I am waiting for a higher level of MOOC technology and methodology. My views and expectations are influenced by my dyslexia and by neuro-diversity.
Note: Waiting does not mean that I do not think about Coursera, edX and Udacity, online learning, policy and MOOCs. One particular issue I keep thinking about is how Coursera, edX and Udacity could kick some status quo a** in Hungary… I wish we had at least one powerful politician saying: “I could give you people free, high quality, global higher education like you have never seen before, so forget about student loans and local narrow minded professors and go global online with us… here is the national plan to make it possible“. I did in fact spend days on the floor drawing colorful charts contemplating this new system, creating a mentor based “university” online with no walls, but with special degrees based on Coursera, edX and Udacity… I do have a plan. Anyway, this was just a side note…
I attempted to ask a female AI researcher in the EU via LinkedIn about testing alternatives, because I started playing with an idea in the human-computer interaction realm, but on a more direct level: wireless brain-computer interaction in learning to eliminate testing and evaluation based stress as much as possible. She requested to connect via LinkedIn (I connected with many Courserians and Coursera employees via LinkedIn) after she read and posted one of my older blog posts. Based on her strong professional background in Artificial Intelligence, I assumed she could answer my new questions, but surprisingly she could not.
My questions are:
– Could learning and teaching happen just as fluidly as a self driving car travels hundreds of thousands of miles on its own?
– Could a computer system take the human brain on a “self driving” learning journey with no stress and with desirably efficient outcomes? I am somehow sure it could, we just need to get there. And for that we first need to ask the right questions.
I came up with these questions while watching the Artificial Intelligence for Robotics lectures by Sebastian Thrun on Udacity.
People of more insight: What do you think about the testing and exam elimination layer? Are there any AI research projects happening already regarding my wish for the elimination of outdated tests and exams? Directly monitoring the brain for clues of learning and knowing would eliminate exam stress, making the entire process of learning so much more fun (and so much healthier… e.g.: I would like to develop a portable monitoring device and smartphone app that tells the learner when to stop and change strategies during learning and examination to keep those damaging stress hormones down, especially in people with learning differences… a bit like a Fitbit Flex, but for your hormones and your brain, for learning and for everyday well-being). Of course I have no problem with essay and project based “measurement” of knowledge (hard to do with a machine, but even that could be more precise with direct brain monitoring… e.g.: we could tell when a feedback is genuine and when it is not, during the Coursera peer assessment process… just need a kit, a headset or a webcam, to monitor the brain or the face of the person giving feedback… (Just a year after I wrote this post I see DARPA starting a neural implant development project to treat chronic diseases by stimulating the peripheral nervous system and read that ENLITIC feeds raw medical data into its AI system for it to learn independently and make its own medical discoveries, diagnoses and later possibly treatments. Now every sceptical person can please stop thinking that my ideas are surreal. We will get to the central nervous system nano implants in no time compared to the age of the universe my friends.)
My wish for a software and interface that could read our brain spikes and interpret the patterns via a computer model as learning, understanding, or confusion and misunderstanding, I think that is related to AI, Coursera (edX, Udacity and other MOOCs), and neuroscience. This could eliminate the unnecessary stress, and change the way we learn, give or gain feedback and produce measurable results in unconscious but well represented and objective, even highly comparative ways.
My readers: Can you think of any team that could already be working on something like this, or would be interested in starting an interdisciplinary project linking “machine learning” to “human learning” so that both components are viewed from a 21st century angle? I am assuming that Coursera’s “machine learning” research project is related to better teaching (as done by the “machine”, not only by humans), but I feel there is the need for an interface that makes it possible for humans to directly tell “the learning machine” where “the learning human brain” is in the “learning process”, objectively and real time.
In the present testing process there is a lot of “noise” that is related to, as I call it: “the indirect interface problem“ (even the “innovative” Coursera testing process is forcing students to adapt to the highly discriminative and rigid testing structure and surface, pushing them to use retro skills related to outdated testing, to be institutionally, neurologically and culturally average to at least pass a test, as if passing a test could be a valid educational goal for the future… well, it is not!), and to the outdated “industrial model of education” (deadlines set by one person for a class of a few dozen, that could be defended in the past… but deadlines set by one person for 80 000 all over the world: hard to defend in the future…). Note: I am aware of the important ethical debate that should be happening parallel to the R&D in (again, as I call it) “brain-machine direct interface learning“. Brainwashing should not be a goal. The ethical process should be about fluid learning.
I am genuinely curious: What do you think about these issues? Also: Are Coursera, edX or Udacity involved in such interdisciplinary AI research? Are “self driving” cars the present and is “self driving” learning the future? And in the very distant future are we going to directly download knowledge into our brains or will we be just taking brain science in a capsule?