|
|
Post by Felicia Ng on Mar 22, 2016 3:49:13 GMT
Such a cognitive view of learning and education honestly concerns me. A LOT. The comment in this thread that strikes and resonates with me the most is Joseph's about what the definition of learning is and what we should actually be striving to impart on students inside or outside the classroom. As he said, if we limit our discussion to the scope of this paper, then sure, cognitive tutors can help improve mathematical cognition and standardized test scores. But emotional growth, cultural awareness, non-linear and non-algorithmic creative skills, and critical thinking abilities are all equally important, if not more important, cognitive developments that need to be addressed in learning and education. Can cognitive tutors ever be designed to address those needs? Where are the limitations to what cognitive tutors can do? Or should do? What even are the societal or philosophical implications of creating a world where teaching is automated by computers?
|
|
Qian
New Member
Posts: 20 
|
Post by Qian on Mar 22, 2016 3:50:16 GMT
I am probably ridiculously misunderstanding the term "cold calculating machines", but what are the emotions in Math? It seems to me Cognitive tutor is not more or less emotional than other edu materials and it seems has no intention to substitute teachers. I think within the context of Math education modeling student comprehension is reasonable and practical. It would be better if the paper described more about the teacher-machine cooperation part.
Following the discussion of evaluation method and standard tests, I'm also wondering if there is a similar mechanism in creativity education. If so, is it possible at all to develop certain kind of *effective* intervention? Even more generally, are the cognition theories apply to creative thinking/education?
|
|
|
|
Post by Brandon on Mar 22, 2016 4:08:12 GMT
Can somebody with more knowledge on cognitive tutors explain to me why there has not been more widespread adoption of this? Improving test scores .3-.4 standard deviations seems like a pretty impressive result. Are more schools actually using these systems or is this still mostly a research thing with a few participating schools? Is it a political will issue? Problems with scalability? Or am I just out of the loop and this is what schools are doing?
I always find the criticism of "cold calculating machines" argument pretty obnoxious. As someone mentioned earlier, it's not like standardized tests or multiplication tables are exactly warm and engaging. It also seems like this paper makes a point of engaging teachers as part of the process and using the tutor as a supplement to other, more standard teaching practices 3 days of the week.
So what am I missing? Are people just opposed to our robot overlords taking over all the teaching jobs so we're settling for worse results? Is this paper presenting too rosy a picture of cognitive tutors' effectiveness? Or is there just a limited set of skills that can be taught this way (due to complexity or lack of developed systems?)
|
|
|
|
Post by Anna on Mar 22, 2016 4:14:12 GMT
Hm, I find this discussion board discussion much more interesting that the paper.
Upshot: I'm thinking about the methods/procedures for the Name Game study, and how limited and messy that whole process sounded. So on the one hand we can think of cognitive tutors (and other, related HCI edu technologies) as able to draw on existing bodies of psychological/social literature. But on the other hand, HCI/cognitive tutors could directly contribute to psychological/behavior research, as they provide a great venue big data analysis that wasn't possible before. (Cheesy? True, though)
|
|
|
|
Post by mrivera on Mar 22, 2016 4:26:32 GMT
@amy's response- Is it clear that those qualities of the language used are intended to act like a human specifically? Perhaps the snark and teasing is more a way to create an external challenger for the student to go up against. It's a lot more enjoyable to prove someone or something wrong than it is just learn something. Both seem to occur with the snark+teasing-based cognitive tutors.
@judith's Bit: "Like everything in HCI - the answer to how these tutors can support students needs is very dependent on a variety of factors and I don't think they do nearly as good a job supporting this diversity." What's interesting is that if you take the perspective of Bjork- the act of experiencing learning in different contexts / environments supports learning. I wonder how this relates to a student who is learning from a cognitive tutor or instructor that does not cater a diversity of language or background. Does exposure to a less than relatable CT promote more effective learning?
|
|
|
|
Post by rushil on Mar 22, 2016 8:25:05 GMT
Someone earlier asked the question, how does it make students feel to become smarter but not feel a sense of reward in the end? The answer is simple. It makes them work harder and push harder. Amy pointed out that there has been research to show students respond better to tutors that don't use polite language. It's not too farfetched to imagine that a strong "keep pushing" environment makes students push boundaries of their potential. This may not be the case with everyone, and everyone won't feel as strongly motivated, but that is true with every form of motivational / persuasive medium. They work and don't work at the same time. This makes me push back and ask the question, why do we need to mold the tutor or any form of aid etc. and how do you determine what is best suited for a particular individual.
|
|
|
|
Post by julian on Mar 22, 2016 18:16:43 GMT
First time reading about cognitive tutors! and I really enjoyed it.
This is a very nice introduction to Cognitive tutors and it left me curious and with a lot of questions about the specific details of how they work and the adaptive strategies it seems to be using for teaching students. Like Judith I'm also interested on how different cultural backgrounds affect whether a Cognitive tutor could even work. From the results it seems like the Cognitive tutor works better than regular courses for students with limiter english proficiency however it would be interesting to dig deeper on what the differences are with respect to native english speakers, and whether there are any patterns on how students are getting confused or stuck.
I found Steven D. comment interesting: "Computers (and robots) are frequently viewed (and portrayed) as cold calculating machines. Human teachers on the other hand of thinking, feeling beings who can detect and adapt to the social and emotional needs of a student". I do not think that human teachers are necessarily like that, in fact from my high school and college experiences, I had way more "robot" professors (cold calculating humans) than the warm social adaptable and nice decent human beings described above. Many of my courses were dull, repetitive and many of the professors would just read from a book or a piece of text, when asked difficult and interesting questions by the students they would often ridicule them in front of the class. Now, I was definitely not in the worst high school or the worst college in my country, these were both top mid tier. Reading this cognitive tutor paper actually reminded me of a software I used in college to learn electronics and how fond I was of it: 0 judgement, I could go back an forth as much as I wanted, it had enough information and it was actually fun. This software, didn't even have a Cognitive Tutor on it and it was awesome!. It is interesting to see how we associate all this positive and nice characteristics of being human when in reality a human can be just as cold as a robot on purpose. Now on the more positive side, I remember fondly also the nice, socially smart and thoughtful professors I studied with, they were fewer than their counterpart but they really left good impressions and guide me all the way to CMU.
Now going back to the cognitive tutors paper, I really liked how throughout the paper is implied that it is not about replacing teachers, it is about making education better and Cognitive Tutors can help with things like designing better curriculums or being used together with regular classroom classes .
|
|
