tag:blogger.com,1999:blog-1245039520249139173.post3847602571906522751..comments2023-11-05T01:02:08.901-07:00Comments on La vertu d'un LA <br>The virtue of an A - A fortunate hive: Information overload - And no more trivia, fool!Laurent DUVALhttp://www.blogger.com/profile/05343286920474971263noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-1245039520249139173.post-22044958882038922372010-06-11T20:43:22.939-07:002010-06-11T20:43:22.939-07:00I remember we already discussed that topic (and we...I remember we already discussed that topic (and we probably should discuss again). There is a large huge gap to fill between "how can I get more data ?" and "is there some oil there ?". The "more data" answer is probably related to a <a href="http://laurent-duval.blogspot.com/2008/05/conferences-concern-fees-about-hammers.html" rel="nofollow">"hammer-like"</a> line of reasoning.<br /><br />Additional questions are:<br />-"what should i have extracted (as useful information) from these data?"<br />-"how useful/harmful (in time or cost for instance) was my acquisition process in reaching the previously mentioned goal?".<br />I wonder for instance whether getting used to 32-bit data somehow fool your inner filter in believing that the 5th digit of outcomes always matters? I will come back later in early "low-fi" sensing in seismics, an early ancestor of <a href="http://nuit-blanche.blogspot.com/" rel="nofollow">compressive sampling</a>. After all, after 40 years of mistreatments and abuses, we did not really learn a lot the Lenna. And not so much from image features in general, isn't it? Yet, we have a huge collection of hammers (to denoise, compress, segment)Laurent DUVALhttps://www.blogger.com/profile/05343286920474971263noreply@blogger.comtag:blogger.com,1999:blog-1245039520249139173.post-45414031218162957732010-06-11T02:10:06.565-07:002010-06-11T02:10:06.565-07:00Laurent,
I think I mentioned that to you a while ...Laurent,<br /><br />I think I mentioned that to you a while back. For instance in the oil business, the question seems to be "how can I get more data ?" while the real question is "is there some oil there ? and how much will it cost me to extract it ?"<br /><br />This line of questioning seems trivial on its own, by that I mean that most engineers will tell you that they want more data to find oil so that they are answering the second question by answering the first, but I sometimes wonder if asking the right question does not lead you to a different trade-off. A little like Clay, if getting petabytes is your everyday problem, maybe your filter has broken somewhere down the line for structural reasons.<br /><br />Cheers,<br /><br />Igor.Igorhttps://www.blogger.com/profile/17474880327699002140noreply@blogger.com