ขนาดวิดีโอ: 1280 X 720853 X 480640 X 360
แสดงแผงควบคุมโปรแกรมเล่น
เล่นอัตโนมัติ
เล่นใหม่
A very beautiful trick! Thanks, man.
Can you recommend any good material for us to ready more about this trick? Book, tutorial, etc? Thanks for the video! Very good!
in the equation ...=b+log {sum {exp(ai-b) } }, what if you use b=min(ai) instead? since b is very negative, ai-b will be positive so underflow won't happen? for example log(exp(-20) + exp(-2)) = -20+log(1+exp(18)) and no underflow here
But if a_i are very big, positive numbers then doing b = max(a_i) prevents the overflow from happening.
Why wouldn't we shift so that 0 is at halfway between the min and the max a_i? Is it because we only care about the largest one?
Cool trick, I didn't know about it.
what drawing software do you use? I would also like to use it for my online class , thanks
Great Video!
really smart trick. Now at least we get a 'b'. ^^
A very beautiful trick! Thanks, man.
Can you recommend any good material for us to ready more about this trick? Book, tutorial, etc? Thanks for the video! Very good!
in the equation ...=b+log {sum {exp(ai-b) } }, what if you use b=min(ai) instead? since b is very negative, ai-b will be positive so underflow won't happen? for example log(exp(-20) + exp(-2)) = -20+log(1+exp(18)) and no underflow here
But if a_i are very big, positive numbers then doing b = max(a_i) prevents the overflow from happening.
Why wouldn't we shift so that 0 is at halfway between the min and the max a_i? Is it because we only care about the largest one?
Cool trick, I didn't know about it.
what drawing software do you use? I would also like to use it for my online class , thanks
Great Video!
really smart trick. Now at least we get a 'b'. ^^