|
Members in Shoutbox
None.
Shoutbox Search
Shoutbox Commands
/w [name] > Whisper
/r > Reply to last whisper /me > Marks as action Shoutbox Information
Moderators may delete any and all shouts at will.
|
Global Shoutbox
Please log in to shout.
[2017-3-30. : 8:14 pm]  lil-Inferno -- I can't think of a single thing that was fixed. The only change I like is the addition of observer mode and even that breaks a few minutes in.[2017-3-30. : 8:08 pm] lil-Inferno -- Yeah, there are so many little bugs. It just reeks of unprofessionalism.[2017-3-30. : 8:04 pm]  Dem0n -- You're literally objectively wrong. They broke far more than they "fixed".[2017-3-30. : 8:03 pm] Dem0n -- ButchButch shouted: existing maps meaning euds basically... which wouldve been the case with any patch. the community was weak af. none of the changes are outright terrible, jus some annoying. they fixed more than they ruined imo and it hasnt been that long. what have they fixed?[2017-3-30. : 8:02 pm] outlawpoet -- ButchButch shouted: existing maps meaning euds basically... which wouldve been the case with any patch. the community was weak af. none of the changes are outright terrible, jus some annoying. they fixed more than they ruined imo and it hasnt been that long. korean community wasn't[2017-3-30. : 8:00 pm]  Ahli -- Dem0nDem0n shouted: tbh I'm confused as to why they released the PTR one day before the expected release. It's pretty clear that like nothing was tested, so I don't know what they expected in terms of how much shit is broken. heroes' team barely fixes bugs within a week of PTR... 1 day PTR means nothing is going to be fixed until release...  [2017-3-30. : 8:00 pm] Butch -- existing maps meaning euds basically... which wouldve been the case with any patch. the community was weak af. none of the changes are outright terrible, jus some annoying. they fixed more than they ruined imo and it hasnt been that long.[2017-3-30. : 7:58 pm] Dem0n -- tbh I'm confused as to why they released the PTR one day before the expected release. It's pretty clear that like nothing was tested, so I don't know what they expected in terms of how much shit is broken.[2017-3-30. : 7:53 pm] l)ark_ssj9kevin -- You actually can sort by game type by just typing it in LOL[2017-3-30. : 7:51 pm] outlawpoet -- ButchButch shouted: the new UI is dope, they jus needa fix a few things like demon said. people are so fuckin critical, im so happy sc is getting any attention at all. We even have an opportunity to suggest shit, 1.18 is not terrible imo unfortunately the attention it's getting is detrimental to existing maps, servers, and communities. It's not like any attention is a good thing[2017-3-30. : 7:51 pm] l)ark_ssj9kevin -- I am a simple man, I see fullscreen windowed mode support, I am pleased[2017-3-30. : 7:50 pm] outlawpoet -- ZoanZoan shouted: Also we are probably mostly just wary because of the shitstorm from SCII and countless other things they fucked up in BW  [2017-3-30. : 7:49 pm] Zoan -- Also we are probably mostly just wary because of the shitstorm from SCII[2017-3-30. : 7:46 pm] Zoan -- ButchButch shouted: the new UI is dope, they jus needa fix a few things like demon said. people are so fuckin critical, im so happy sc is getting any attention at all. We even have an opportunity to suggest shit, 1.18 is not terrible imo It needs to look different than the game start lobbies too[2017-3-30. : 7:46 pm] TiKels -- so if you were able to calculate it in an amount of time represented by a polynomial distribution with 100% accuracy it would be considered a P solved problem?[2017-3-30. : 7:45 pm] Butch -- the new UI is dope, they jus needa fix a few things like demon said. people are so fuckin critical, im so happy sc is getting any attention at all. We even have an opportunity to suggest shit, 1.18 is not terrible imo[2017-3-30. : 7:41 pm] TiKels -- Do you just determine that the number of cities (I assume that's the input?) doesn't leave a solution where the time is listed something like 2^n (where n is number of cities)?[2017-3-30. : 7:40 pm] TiKels -- I get the rough idea, but how do you determine whether something is calculated under "Polynomial time"?[2017-3-30. : 7:39 pm] TiKels -- I feel like you're talking on a level way above my current knowledge haha[2017-3-30. : 7:38 pm] Dem0n -- text box doesn't stay focused so you have to click on it every time you wanna type if you switch windows[2017-3-30. : 7:38 pm] Dem0n -- escape no longer clears your message and instead quits multiplayer, alt+c is no longer "create" and is instead "cancel" (ie quit multiplayer)[2017-3-30. : 7:35 pm] outlawpoet -- that's why we make approximation algorithms that get close enough[2017-3-30. : 7:34 pm] outlawpoet -- AI isn't some panacea, it still relies on CPU time for calculations. An arbitrarily large TSP isn't going to be solvable in polynomial time regardless of the method[2017-3-30. : 7:32 pm] outlawpoet -- TiKelsTiKels shouted: But you would never know if you always got correct answerst with AI. It could give you 10000 optimal answers to 10000 problems, but you'll never know if there is an example of the problem it cant solve in the TSP it's not like an AI is going to have access to different information or better algorithms. What it might do better than humans is choose which algorithm is going to result in the best approximation for any given subset of the data[2017-3-30. : 7:31 pm] lil-Inferno -- Tbh fam, w/e new UI, people should be mad AF that they removed features and fucced up others LMAO[2017-3-30. : 7:26 pm] Ahli -- AhliAhli shouted: to detect a structure being under construction, we center a location on that structure type, right? then it centers on it and does not include a certain unit anymore that would be included in the center of the map, right? can we focus on the real problems... that was the right approach to fastest way to detect supply depot construction, right?  [2017-3-30. : 7:23 pm] outlawpoet -- TiKelsTiKels shouted: But you would never know if you always got correct answerst with AI. It could give you 10000 optimal answers to 10000 problems, but you'll never know if there is an example of the problem it cant solve ya, NP-hard problems often take as long to verify as they do to compute[2017-3-30. : 7:22 pm] outlawpoet -- there's a formulation that is NP-complete - "given a length L, decide whether the graph has any tour shorter than L"[2017-3-30. : 7:21 pm] TiKels -- But you would never know if you always got correct answerst with AI. It could give you 10000 optimal answers to 10000 problems, but you'll never know if there is an example of the problem it cant solve[2017-3-30. : 7:21 pm] outlawpoet -- AhliAhli shouted: I would replace the word correct with optimal... correct would be any trip that manages to reach all customers. But you want the optimal plan and not just any... depends on the formulation of the TSP. typically framed as shortest possible distance / least expensive route[2017-3-30. : 7:20 pm] Ahli -- I would replace the word correct with optimal... correct would be any trip that manages to reach all customers. But you want the optimal plan and not just any...[2017-3-30. : 7:19 pm] TiKels -- Outlaw poet wanna take this to a discord or something? You seem pretty knowledgeable[2017-3-30. : 7:19 pm] outlawpoet -- ya, it just takes a long ass time. NP stands for non-deterministic polynomial time[2017-3-30. : 7:16 pm] Dem0n -- TiKelsTiKels shouted: outlawpoet outlawpoet My question is if you designed an algorithm that was able to efficiently solve the traveling salesman problem but were not able to demonstrate that it always got the correct solution (without brute forcing), while being (for lack of a better word) exponentially faster... is that not a solution to an NP problem? Sorry if my understanding of NP is a little elementary If you can't prove 100% correctness, it's not a solution.[2017-3-30. : 7:15 pm] TiKels -- outlawpoetoutlawpoet shouted: I'm not sure I follow your question, TiKels. If you get the correct solution every time, either your heuristic/algorithm is extremely good, extremely lucky, or your search space is tiny outlawpoetoutlawpoet shouted: outlawpoet or you spent a fuckton of CPU time calculating it LOL. solutions are inevitable, mostly it's about time to solve My question is if you designed an algorithm that was able to efficiently solve the traveling salesman problem but were not able to demonstrate that it always got the correct solution (without brute forcing), while being (for lack of a better word) exponentially faster... is that not a solution to an NP problem? Sorry if my understanding of NP is a little elementary[2017-3-30. : 7:14 pm] Dem0n -- l)ark_ssj9kevinl)ark_ssj9kevin shouted: lol there's a fps cap gotta make sure it doesn't get too out of hand. some computers may not be able to handle it![2017-3-30. : 7:14 pm] outlawpoet -- answers to* calculations. I don't care how good your genetic algorithm is, you're not gonna solve a million city TSP without a fuckton of CPU time[2017-3-30. : 7:12 pm] outlawpoet -- TiKelsTiKels shouted: How do AI techniques fit into the P vs NP paradigm? ok I think I understand your angle better now. AI doesn't magically produce answers calculations that can't be done in polynomial time, but it could come up with good heuristics for doing so[2017-3-30. : 6:59 pm] Ahli -- to detect a structure being under construction, we center a location on that structure type, right? then it centers on it and does not include a certain unit anymore that would be included in the center of the map, right?[2017-3-30. : 6:58 pm] outlawpoet -- outlawpoetoutlawpoet shouted: I'm not sure I follow your question, TiKels. If you get the correct solution every time, either your heuristic/algorithm is extremely good, extremely lucky, or your search space is tiny or you spent a fuckton of CPU time calculating it LOL. solutions are inevitable, mostly it's about time to solve[2017-3-30. : 6:58 pm] Azrael -- It's about how much the UI sucks, and the developer basically says "we didnt like the old UI so we made it the way we like it, you better get used to it lol". |
