3 ms·
Suggestion: filter the titles that are identical to real ones. Or maybe it was part of the plan, demonstrating how they are indistinguishable. My deadly di
by BoppreH 13y ago
Suggestion: filter the titles that are identical to real ones. Or maybe it was part of the plan, demonstrating how they are indistinguishable.
My deadly disease was just a website.
IBM's "Watson" finally ready for preorder.
TileMill — an application for making my daughter cry.
"Anonymous" Hackers Take Down Megaupload Video
Google finally has a natural, almost visceral aversion to complexity
Syrian Internet Is a Hero
IRS claims it does not compute
Stripe raises $18 million from affiliate links on "Ad-Free" blog
Why I use paper to wipe my butt every day.
Apple not providing LGPL webkit source code 23 years later and you're still loving
- gamegoblin 13y agoIf you look at the code, I filter all titles that are substrings (including identical) of real ones. Sometimes the Markov chain will simply add a single word to it, though, so it gets through the filter. Any suggestions on how to improve it?
- BoppreH 13y agoFunny, I'm sure I google'd a few of them and got a perfect match, but now I can't find any more. Probably my fault, sorry.
- sillysaurus2 13y agoThe straightforward solution is: For each fake title, compute its Hamming distance to every real title. Filter titles whose distances are very close to real titles. It's O(N^2), but if you write it in Python and run it with pypy, performance may be acceptable. Maybe. But I think this is a classic example of a Problem That Doesn't Matter. No need to spend time solving it (unless you're having fun!)
- kps 13y agoAnother suggestion: terminate unmatched open delimiters at the end, and strip unmatched close delimiters (parentheses, quotation marks, etc.