Teachers Aren’t All Made from Meat

Everybody can agree that the number of teachers versus students significantly effects the quality of education. With a high number of students to each teacher, the courses have to be adapted to the group and less attention payed to each students characteristics, strengths and weaknesses. I’ve been aware of this pesty fact through my own experience of school and consider it a noteworthy problem of modern education. So, let’s mass produce teachers to come to the rescue.

Montreal based company uMind made news yesterday with their AI eLearning software. Claude Frasson, founder of uMind, describes the system as an artificially intelligent tutor that actively evaluates the emotional distress, weaknesses and strengths of the student. Using this information as a basis, it custom-builds course material as the student advances. Taking a look at their webpage header, you’ll notice that education software is a looker as well. A virtual humanoid named Aimy. Via the Gazette:

“Intelligence means adapting to the learner and understanding the capability of the learner,” said Frasson, founder and president of uMind. “In any exam, you have a lot of learners who are able to succeed. We have found all the mechanisms (that help students succeed) and we have applied them to eLearning.”
Launched in March, uMind employs AI to create a virtual tutor that recognizes and adapts to the student’s limitations and emotional distress. The instructor knows when a student is stumped and activates extra teaching modules on the specific subject.

But have we reached a point where this kind of software actually does any good? Does it rival us meatmachines? Not really. But Frasson estimates that this approach, when compared to traditional eLearning, cuts the learning time in half and increases the student retention by 35%. The program also includes a virtual reality type mode, which has been used by the Montreal Transit Company to train employees in handling emergencies. Early estimates indicate training costs close to being reduced by 50%. (That’s according to this one article and man, I haven’t properly confirmed it.)

Hannes Vilhjalmsson over here at RU’s AI lab CADIA has worked on similar virtual environments for education in the past. His Tactical Language and Culture Training systems (TLCTS, developed at ISI) are aimed at training language skills and have, for example, been used by the US military to train soldiers in speaking Iraqi and Pashto.

From his webpage:

The objective of these series of projects was to develop tools to support individualized language learning, and apply them to the acquisition of the linguistic, gestural, and cultural knowledge and skills necessary to accomplish specific tasks in a foreign environment.

In short, the student plays a “serious game” where, in order to sucessfully accomplish his missions, he has to interact verbally with artificially intelligent agents. For those interested, these projects use the open source Unreal game engine. Really cool stuff, and you can imagine how much I wish I had this kind of system to train me german and danish in my younger years.

No related posts detected.

5 Comments, Comment or Ping

  1. “For those interested, these projects use the open source Unreal game engine.”
    … Sweet.

    Personally, Id love to learn Spanish thru GTA type of game setting. The whole idea is genius, especially in regards to learning languages.

    This needs to be brought to the public. Someone needs to contact RockStar quick!

  2. Hey John — yeah, this type of software and research is definitely worth persuing.

  3. First of all, in the corporate world I had a LOT of people trying to sell me similar training systems for my staff at great cost.

    Every one of them claimed it could improve retention and comprehension, and all incorporated what they claimed were AI type components. However, there are opposing views to all of this marketing hype.

    – The machines cannot guage a human’s interest in a subject, nor can they do anything to increase it. Great teachers can.

    – The cost of such machines are extraordinary. You could hire more teachers for the same money.

    – The programs and hardware need to be updated routinely which means there are ongoing costs in addition to the upfront costs.

    If you need any proof that skepticism is in order just examine this one statement:

    “We have found all the mechanisms (that help students succeed) and we have applied them to eLearning.”

    Oh really? So, how does the system help when a person is feeling a little ill, or worried about a relative in the hospital, or nervous about an upcoming test, or tired?

    eLearning is an excellent tool to augment traditional learning methods, but attempting to replace them altogether is absolute folly.

    John

    PS – I have helped to develop eLearning programs so I have some experience in this area.

  4. Hey John.
    Yeah, of course this system doesn’t replace a human, and that Gazette article is definitely exaggerating its potential — like the media so often tends to do.

    But it might make up for a lack of human teachers, provided they are actually cheaper.

    I get the feeling you think I considered this a replacement for humans? That was not my intention — as I understood the article, they are comparing their AI system to “static” or non-interactive eLearning.

    That is, a replacement for traditional eLearning tools: reading static text and answering exams versus having an AI agent help you study.

    I just realized reading over the article that the “But does it rival us meatmachines?” implies this, woops. My bad. I added it as “extra spice” when I was done writing the entry and didn’t realize until now that I hadn’t explained this was a comparison of two eLearning methods — not AI and human teachers.

    Fixed now. Thanks.

  5. By the way, can you give me a price range for these kinds of systems? Just an estimate, based on your experience?

Reply to “Teachers Aren’t All Made from Meat”

Please read the Terms of Use before commenting!

Basic HTML allowed (a, blockquote, strong, em)


Other ..

Think Artificial is a proud member of the 9rules blog community.