TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

A Horrible Experiment

87 点作者 marvindanig超过 4 年前

15 条评论

kingosticks超过 4 年前
There&#x27;s a whole lot of random people chiming in and guessing on Twitter (shocker) so it is worth pointing out that:<p>&gt; The algorithm does not do face detection at all (it actually replaced a previous algorithm which did).<p>From <a href="https:&#x2F;&#x2F;twitter.com&#x2F;ZehanWang&#x2F;status&#x2F;1307461285811032066?s=20" rel="nofollow">https:&#x2F;&#x2F;twitter.com&#x2F;ZehanWang&#x2F;status&#x2F;1307461285811032066?s=2...</a><p>Not to say there is or isn&#x27;t a problem to be fixed, more work is needed to understand. Such as the analysis at <a href="https:&#x2F;&#x2F;twitter.com&#x2F;vinayprabhu&#x2F;status&#x2F;1307460502017028096?s=20" rel="nofollow">https:&#x2F;&#x2F;twitter.com&#x2F;vinayprabhu&#x2F;status&#x2F;1307460502017028096?s...</a>
评论 #24534109 未加载
dazbradbury超过 4 年前
<a href="https:&#x2F;&#x2F;twitter.com&#x2F;SergioSemJ&#x2F;status&#x2F;1307493041742254080?s=19" rel="nofollow">https:&#x2F;&#x2F;twitter.com&#x2F;SergioSemJ&#x2F;status&#x2F;1307493041742254080?s=...</a><p>One theory is it&#x27;s actually the glasses increasing the likelihood of correctly identifying a face (or &quot;feature&quot; element in the photo). Interesting looking through the various tests in the thread, as clearly lots of variables here, but no way this will be exhaustive.<p>Suspect this might kick off some serious arguments at Twitter HQ for allowing users to pick the cropping vs. an algorithm regardless of any bias found one way or not.
评论 #24541222 未加载
DanBC超过 4 年前
See also Zoom backgrounds: <a href="https:&#x2F;&#x2F;twitter.com&#x2F;colinmadland&#x2F;status&#x2F;1307111816250748933?s=20" rel="nofollow">https:&#x2F;&#x2F;twitter.com&#x2F;colinmadland&#x2F;status&#x2F;1307111816250748933?...</a>
TazeTSchnitzel超过 4 年前
Wonderful example of the ethics of ML. The people whose eyes they tracked gravitated towards certain kinds of faces (not necessarily just based on skin colour), so now Twitter&#x27;s model crops other faces out of pictures, among other problems (now it likes cleavage apparently).<p>The old approach may have been a bit mechanical but it avoided the non-neutrality of unconscious human perception. The assumption that what human eyes gravitate to is what an image should be cropped to is a very big and questionable one. The first thing I look at in a picture is not always the most important part, and the image as a whole has value.
noxer超过 4 年前
The real problem I see here is the default assumption from many people that these is racial bias in the algorithm with no evidence thereof whatsoever. Thats how twitter works, shitstorm first, research later, evidence or corrections for false assumptions - never.
评论 #24534261 未加载
评论 #24534257 未加载
fr2null超过 4 年前
Interestingly, swapping the glasses does seem to make the thumbnail Obama. [1]<p>1. <a href="https:&#x2F;&#x2F;twitter.com&#x2F;SergioSemJ&#x2F;status&#x2F;1307493041742254080?s=19" rel="nofollow">https:&#x2F;&#x2F;twitter.com&#x2F;SergioSemJ&#x2F;status&#x2F;1307493041742254080?s=...</a>
评论 #24533996 未加载
greenduck超过 4 年前
This is pretty egregious.<p>For those of us working on machine learning, does anybody know of datasets that have decent representation of different colors of skin. That seems like the cause here and thinking back, my datasets aren&#x27;t the most diverse out there.
评论 #24534042 未加载
评论 #24536036 未加载
评论 #24534391 未加载
pityJuke超过 4 年前
Some more info:<p>- This apparently done via a NN [1]<p>- There are mixed results, with various factors playing a factor [2].<p>Worth digging into this thread, and the original thread where this was figured out. It&#x27;s intriguing.<p>[1]: <a href="https:&#x2F;&#x2F;blog.twitter.com&#x2F;engineering&#x2F;en_us&#x2F;topics&#x2F;infrastructure&#x2F;2018&#x2F;Smart-Auto-Cropping-of-Images.html" rel="nofollow">https:&#x2F;&#x2F;blog.twitter.com&#x2F;engineering&#x2F;en_us&#x2F;topics&#x2F;infrastruc...</a><p>[2]: <a href="https:&#x2F;&#x2F;mobile.twitter.com&#x2F;NotAFile&#x2F;status&#x2F;1307337294249103361" rel="nofollow">https:&#x2F;&#x2F;mobile.twitter.com&#x2F;NotAFile&#x2F;status&#x2F;13073372942491033...</a>
t0astbread超过 4 年前
Why does Twitter even try so hard to make preview crops &quot;meaningful&quot;? Sure, it produces interesting results sometimes (sometimes funny, sometimes less so) but isn&#x27;t it a lot of work for not a lot of benefit?
评论 #24534355 未加载
评论 #24534321 未加载
评论 #24534230 未加载
nailer超过 4 年前
Misleading - it seems to be looking for contrast<p><a href="https:&#x2F;&#x2F;twitter.com&#x2F;kim&#x2F;status&#x2F;1307548258491801600?s=19" rel="nofollow">https:&#x2F;&#x2F;twitter.com&#x2F;kim&#x2F;status&#x2F;1307548258491801600?s=19</a><p>UPDATE: someone wrote a boy to test the conspiracy and properly debunked it, source code is online:<p><a href="https:&#x2F;&#x2F;twitter.com&#x2F;vinayprabhu&#x2F;status&#x2F;1307497736191635458?s=19" rel="nofollow">https:&#x2F;&#x2F;twitter.com&#x2F;vinayprabhu&#x2F;status&#x2F;1307497736191635458?s...</a>
daenz超过 4 年前
Isn&#x27;t Mitch McConnell in the news right now far more &quot;intensely&quot; than Obama? If I was writing a system to pick a thumbnail, I would pick the thumbnail that is most relevant to current tweets, which is definitely McConnell.<p>A quick way to test this would be to use two well known figures, where the PoC is definitely more in the news than the other.
dmitriid超过 4 年前
It&#x27;s funny how these big companies willingly convert a non-problem into massive problems with hundreds of hours of engineering effort and huge social implications.<p>Non-problem: don&#x27;t mess with the images, and display them as is.<p>Problem: see Twitter
rambojazz超过 4 年前
I&#x27;m out of the loop. What&#x27;s happening here? All I see is a picture.
评论 #24541140 未加载
Gys超过 4 年前
It seems this needs a twitter account to see? Maybe someone can post a screenshot?
评论 #24533927 未加载
Kednicma超过 4 年前
Humans can&#x27;t make unbiased computers. Every computer which humans make is indelibly human as well. This is the anthropic bias.
评论 #24534283 未加载