Baby Meets Big Brother For Science 188
dylanduck writes "A baby is to be monitored by a network of microphones and video cameras for 14 hours a day, 365 days a year, in an effort to unravel the seemingly miraculous process by which children acquire language. I guess that's what happens when your pop works at MIT's Media Lab. Thankfully his parents can switch off the surveillance for 'private' moments and delete short scenes. All the footage is being classified by algorithms."
Videos (Score:3, Interesting)
If you can't see them, there are 9 fish eye cameras mounted at certain points of the house and a day passes in 30 seconds (a la National Geographic plant blooming or Requiem for a Dream old lady on crack).
Each camera seems to have a round piece of paper ready to flip up and down to cover it (possibly via light switch in the room/area) should the family choose it to be necessary.
I think this is a wonderful and innovative idea, my only concern resides in the child's rights. I'm going to say I don't agree with even releasing these short clips to the public. I believe that this footage should be collected, protected & anonymity of the child enforced until the child is 18--at which point they will be capable of releasing the footage under whatever license (GPL even, lol) they deem appropriate. I understand that the parents have full custody, I only hope this child is in no way taken advantage of like so many prodigious children are by their parents.
Computational Linguistics (Score:3, Interesting)
What would be better is to develop algorithms that try and mimic the learning process we already have observed in native language acquisition and then continue to refine our algorithms until we have perfected that process. We will only know we have it right when you can take those same algorithms, put them to use by exposing it to a different language and have it still learn it right.
I did something like this (Score:4, Interesting)
What I found is that the sample size was way too small. Almost every child has vastly different development patterns and to see the big picture you need a bigger sample than one kid. We're talking about a huge effort to collect that much data on many children but I think that is what will be required to even begin to understand how it works.
Ever notice that what is easy is hard? (Score:3, Interesting)
Just like it is easy to write a program that can calculate sin but really hard to write one that can follow verbal directions as well as a a four year old?
In other words it is easy to teach a machine what is taught in school.
It is very hard to teach a machine what is taught by parents.
Re:Videos (Score:3, Interesting)
the 18 year old can still have regrets later...
Re:Proud first words (Score:4, Interesting)
There's one thing I still don't understand. Sure, I know that McDonalds all pervasive advertising campaign virtually assure brand recognition by age three. I know that McDonalds isn't the only company engaged in this. I know how this works.
No. What I don't understand is why it works. Why do children fixate on McDonalds so much? What is the secret sause here? And it's not just McDonalds. Apparently, brand loyalty can be instilled before the third year.
This perplexes me. What's driving these kids to say McDonalds before Mommy?