Confession: What makes me good at machine learning research makes my life miserable. And that is: I habitually stick with a problem until it's solved. The problem is that the problem can never truly be solved in machine learning. 1/
-
Show this thread
-
Engineering problems, in contrast, are generally finite. That's fun. The obsessive problem solving has a natural end: When the problem is solved. Two years of having an unsolved problem with this type of personality? That SUCKS. Now I know. 2/
2 replies 0 retweets 25 likesShow this thread -
I plan on writing more about this stuff later- about to wrap up research with a hard deadline. It's going to be a relief I think. Then what? Time to get a life again. 3/
5 replies 0 retweets 32 likesShow this thread -
Replying to @citnaj
Hi Jason, do you know some criterias or rule of thumb to judge whether a ML problem is solvable or not?
1 reply 0 retweets 0 likes
Well by solvable I mean reaching a perfect solution where there's no doubt left- it's doing everything right. So I'm really alluding to the problems this presents to a perfectionist. I really don't think colorization can be truly solvable in that sense.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.