tag:blogger.com,1999:blog-9275314.post2533094246288225187..comments2024-03-23T11:34:23.998-04:00Comments on Togelius: The differences between tinkering and researchJulian Togeliushttp://www.blogger.com/profile/09333191187316058782noreply@blogger.comBlogger10125tag:blogger.com,1999:blog-9275314.post-8620076246648744372019-02-06T16:14:10.915-05:002019-02-06T16:14:10.915-05:00By calling him "Some guy" repeatedly and...By calling him "Some guy" repeatedly and never using his username or given name, it very much comes off as criticism and it did to me as well. Thank you for clarifying in the comments that it isn't that, but I got exactly the same feeling from this that Danny did due to diction.<br /><br />Besides that, the article is helpful and interesting. Thank you.Andrewhttps://www.blogger.com/profile/15403896855662881572noreply@blogger.comtag:blogger.com,1999:blog-9275314.post-2539108389497263132018-04-02T23:42:00.714-04:002018-04-02T23:42:00.714-04:00@Danny: He mentions NEAT, the algorithm he uses, a...@Danny: He mentions NEAT, the algorithm he uses, and that's good. But that's the only work he mentions. He does not mention any approaches to playing Mario, or even solving a similar problem. A real discussion of previous work would look at different approaches to this type of problem, and their pros and cons. And if there's someone who has done essentially the same thing before, that should be mentioned...<br /><br />Note that I'm not even criticizing SethBling here. He did a great job with that video, and there's no law that you need to have a related works section in your YouTube videos. Having it would probably have made the video less engaging. I'm just using him as an example of when something is not research, because of the lack of scholarship.Julian Togeliushttps://www.blogger.com/profile/09333191187316058782noreply@blogger.comtag:blogger.com,1999:blog-9275314.post-9517323971202785232018-04-02T23:18:16.103-04:002018-04-02T23:18:16.103-04:00Looks like you didn't care to watch the video....Looks like you didn't care to watch the video. From 4:40 the guy makes plenty of reference to prior work, names authors of papers and calls out the name of the methodology used.<br /><br />https://youtu.be/qv6UVOQ0F44?t=4m40s<br /><br />Your statement "The guy didn't even know he was reinventing the wheel, and didn't care to look it up." is completely and blatantly false, making your criticism very weak.Dannyhttps://www.blogger.com/profile/05651557455925895353noreply@blogger.comtag:blogger.com,1999:blog-9275314.post-15852345133479457012016-05-25T18:44:48.433-04:002016-05-25T18:44:48.433-04:00@hermal: in biology you would have a good point, b...@hermal: in biology you would have a good point, but in computer science (especially AI / machine learning) the papers are open to the public and not hidden behind paywalls. Alhttps://www.blogger.com/profile/17327836181389523648noreply@blogger.comtag:blogger.com,1999:blog-9275314.post-75158264189945859552016-05-20T07:33:48.332-04:002016-05-20T07:33:48.332-04:00One thing perhaps of note is that SethBling, i.e. ...One thing perhaps of note is that SethBling, i.e. the youtube mario guy, may have actually contributed something new when he applied the same algorithm to MarioKart (<a href="https://www.youtube.com/watch?v=S9Y_I9vY8Qw" rel="nofollow">see here </a>) which is, as far as I know, a novel learner representation for racing games. Indeed, it's the only instance that I have encountered of the local perception grid style of representation being used for a 3D game. Of course it's entirely possible that I'm just showing my lack of due diligence and that the same thing has been done a thousand times over. <br /><br />Either way (for if I have been lazy, it has at least introduced me to the idea) it goes to show that there is value to be gained from tinkerers (not that I believe that you were suggesting otherwise). Furthermore, I believe the suggestion was given to him by the community, so his latter experiment may never have happened if the idea had remained buried in some paper somewhere.Anonymoushttps://www.blogger.com/profile/12324981280936660390noreply@blogger.comtag:blogger.com,1999:blog-9275314.post-54627711865375064002016-04-21T11:10:02.654-04:002016-04-21T11:10:02.654-04:00Interesting article, however there is one thing mi...Interesting article, however there is one thing missing. The tinkering guy many times does not have access to those fancy papers. Don't get me wrong but this is a fact that people at elite universities have all the subscriptions and people working on their own have less or no access to them. Buying papers is not a gem of solution for the tinkering guy because every topics has tons and tons of peer reviewed papers now. Moreover, I don't know of other fields but computer science guys do not regularly put the code for using their method. The guy who just wants to try the method and use it to build something new has to redo the work anyways because there is no code and the language of the paper is at times very "scientific" and make papers hardly reproducible.As much as I agree with many points that you put forward I think that the tinkering guy would use or reuse. You made some good points there Kudos :-) Hemalhttps://www.blogger.com/profile/03510699163593440490noreply@blogger.comtag:blogger.com,1999:blog-9275314.post-168289277511985892016-04-20T09:37:53.178-04:002016-04-20T09:37:53.178-04:00This comment has been removed by the author.Anonymoushttps://www.blogger.com/profile/13361111646279537101noreply@blogger.comtag:blogger.com,1999:blog-9275314.post-83042663506740679432016-04-19T11:10:00.493-04:002016-04-19T11:10:00.493-04:00Quite a thrilling post! Enjoyed reading...I am mys...Quite a thrilling post! Enjoyed reading...I am myself a tinkerer, doing on and off research on Quantum Computing And AI. Thx for the wonderful article. Stay blessedSHAHhttps://www.blogger.com/profile/06333725644446648542noreply@blogger.comtag:blogger.com,1999:blog-9275314.post-87124269420275975232016-04-18T20:10:33.651-04:002016-04-18T20:10:33.651-04:00Fixed. Not a question about research as much as wr...Fixed. Not a question about research as much as writing speed.Julian Togeliushttps://www.blogger.com/profile/09333191187316058782noreply@blogger.comtag:blogger.com,1999:blog-9275314.post-54422735246256867092016-04-18T01:17:19.466-04:002016-04-18T01:17:19.466-04:00Sir Alexander Fleming discovers benzylpenicillin, ...Sir Alexander Fleming discovers benzylpenicillin, Ian Fleming discovers James Bond. Basic research starts with Wikipedia?Anonymoushttps://www.blogger.com/profile/00648371401005453741noreply@blogger.com