Re:Why Probability Theory Should Be Thrown Under the Bus
by loceng on Friday November 03, 2017 @11:21AM
(#15619829)
Unfortunately most people who haven't become a particular expert in any area don't understand the idea of nuances and how deep they really can go. The more I've become an expert in an area or through multi-disciplinary interests and the nuances between those areas, the more I feel I've become better at differentiating how much to integrate someone's thoughts/understanding based on the level of depth of their arguments - contrasting to how much detail I expect myself to put into an argument if I am trying to prove a certain point. Becoming nuanced is a practice, regurgitating shallow depths of an argument is generally more an state of memorizing - which for me doesn't build trust that a person truly understands something.
Not sure I wrote all this out, one of those reflective mornings..
Re:Why Probability Theory Should Be Thrown Under the Bus
by pjc50 on Friday November 03, 2017 @11:29AM
(#15619895)
The main thing in this is the cool demo of reverse-generating a distribution from its stats to produce lots of very different shapes with statistically identical properties. Like the adversarial images discussion recently.
But I think people over-estimate what can be taken from statistics. It's a tricky subject; ultimately the map is not the territory and condensing a reality to a distribution loses information. And probability can be very counter-intuitive from the Monty Hall problem upwards.
Re:Why Probability Theory Should Be Thrown Under the Bus
by selimthegrim on Friday November 03, 2017 @11:30AM
(#15619904)
I was lucky enough to see a series of lectures by Surya Ganguli this summer - he stole the show at the conference I was at. I highly recommmend seeking out his presentations.
GASP! More clickbait "content marketing".
Not sure I wrote all this out, one of those reflective mornings..
But I think people over-estimate what can be taken from statistics. It's a tricky subject; ultimately the map is not the territory and condensing a reality to a distribution loses information. And probability can be very counter-intuitive from the Monty Hall problem upwards.