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。

Does Machine Translation Affect International Trade?

80 点作者 cracker_jacks超过 4 年前

6 条评论

userbinator超过 4 年前
I thought the "large digital platform" would be AliExpress (which from experience I know does use MT extensively in seller-buyer communications), but it's actually referring to eBay. My experience with the former is that the MT is far from perfect, and even frequently causes much amusement, but definitely useful.
评论 #24615228 未加载
评论 #24611915 未加载
评论 #24611877 未加载
pjfin123超过 4 年前
Shameless plug for my open source neural machine translation project. A Python library and GUI with pre-trained models that packages the files needed for machine translation and makes them usable through a simple interface.<p><a href="https:&#x2F;&#x2F;github.com&#x2F;argosopentech&#x2F;argos-translate" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;argosopentech&#x2F;argos-translate</a>
评论 #24612590 未加载
peteretep超过 4 年前
I’ve got to believe Sichuan food would be more popular if 口水鸡 was less often translated as “saliva chicken” than “mouth-wateringly good chicken”
评论 #24613407 未加载
评论 #24612828 未加载
评论 #24612341 未加载
评论 #24612063 未加载
评论 #24612319 未加载
评论 #24615156 未加载
dwohnitmok超过 4 年前
Machine translation feels like one of the quiet victories of the current crop of machine learning products.<p>It&#x27;s getting to be within spitting distance of being completely adequate (though not excellent) for translating long-form articles, which is amazing considering where things stood ten years ago.
评论 #24612309 未加载
评论 #24612034 未加载
评论 #24613512 未加载
评论 #24612993 未加载
unhammer超过 4 年前
For low-resource languages, or anything of great linguistic distance, neural MT is still not there. That is to say, you may find some language pairs where it works OK, but there&#x27;s a huge sea of pairs where all you get is gibberish – and still they publish these systems and let people use them to produce more gibberish which they put on their blogs and spam sites, making the non-English web into a corpus of dissociated press ramblings.<p>Baidu recently added Saami to their system. They manage to make a Wikipedia article about wolverines into one about spreadsheets, depressed buttons and famous chickens: <a href="https:&#x2F;&#x2F;imgur.com&#x2F;a&#x2F;m91SbRw" rel="nofollow">https:&#x2F;&#x2F;imgur.com&#x2F;a&#x2F;m91SbRw</a><p>A rough translation of what Baidu gave:<p>&gt; No one is greater than zero. There are new things for spreadsheet[here Baidu used the English word for some reason] and all the things in it.<p>&gt; English is an active wolf, but the background as wolf indicates that the active button is depressed.<p>&gt; Too many places to install. Sweden was released from 1968 and July 1973. Probably number of times the alarm is to be reapeated within 39 minutes. 2010 is in total 66 chickens, but later a famous chicken and chickens 12th of march 2010, with a total of 54 chickens<p>Apertium&#x27;s translation is often stilted, but it gets the meaning across: Wolverines are the largest animals in the marten family, etc. Google Translate isn&#x27;t half bad at Norwegian→English, so you can even combine them to get a sort of Saami→English: <a href="https:&#x2F;&#x2F;translate.google.com&#x2F;#view=home&amp;op=translate&amp;sl=no&amp;tl=en&amp;text=Jerven%20(Gulo%20gulo)%20er%20hos%20m%C3%A5rfamilien%20med%20det%20st%C3%B8rste%20dyret.%20De%20andre%20store%20m%C3%A5rdyrene%20er%20Storoter%20og%20havoter.%20Jerven%20finnes%20spredt%20i%20cirkumpolare%20fjell%20og%20h%C3%B8yt%20landskap.%0A%0AP%C3%A5%20engelsk%20er%20det%20%C2%ABjerv%C2%BB%20wolverine%2C%20hvis%20bakgrunnen%20er%20wolvering%20som%20som%20verdien%20det%20betyr%20at%20%C2%ABhan%20oppf%C3%B8rer%20seg%20som%20ulv%C2%BB.%20%0A%0AN%C3%A5%20har%20mange%20land%20fredlyst%20jerven.%20Se%20Bern-konvensjonens%20%C2%ABLista%20II%C2%BB%2C%20som%20ble%20etablert%20av%201982.%20I%20Sverige%20ble%20geatki1968%20fredet%23%3Ai%20s%20og%20S%C3%B8r-Norge%20i%201973.%20Myndighetene%20har%20bestemt%20de%20minste%20tallets%20rovdyr%2C%20som%20skal%20v%C3%A6re%20i%20Norge.%20Jerven%20med%20hensyn%20til%20det%20er%20m%C3%A5lsetting%20at%20de%20skal%20v%C3%A6re%2039%20f%C3%A5%20i%20%C3%A5ret.%20I%202010%20er%20de%20for%20informasjonen%2066%20f%C3%A5%20ciiko%2C%20men%20etter%20jakta%20hvor%20det%20drepte%20ciiko%20og%20ungene%20i%2012%20hi%2C%20s%C3%A5%20var%20de%20v%C3%A5ren%202010%20i%2054%20hi%20unger" rel="nofollow">https:&#x2F;&#x2F;translate.google.com&#x2F;#view=home&amp;op=translate&amp;sl=no&amp;t...</a>.<p>I tried Baidu again now with the same input, just to see what they would give into English, and now it&#x27;s about old ports, finished volcanos and civic computers:<p><a href="https:&#x2F;&#x2F;fanyi.baidu.com&#x2F;#sme&#x2F;en&#x2F;Geatki%20(Gulo%20gulo)%20lea%20stuorimus%20ealli%20neahtebearra%C5%A1is.%20Ear%C3%A1%20stuora%20neahteeallit%20leat%20Stuora%C4%8Deavrris%20ja%20meara%C4%8Deavrris.%20Geatki%20g%C3%A1vdno%20bie%C4%91gguid%20sirkumpol%C3%A1ra%20v%C3%A1riin%20ja%20alla%20eatnamiin.%0A%0AE%C5%8Bgelasgillii%20lea%20%C2%ABgeatki%C2%BB%20wolverine%2C%20man%20duog%C3%A1%C5%A1%20lea%20wolvering%20mii%20%C3%A1rvvusge%20mearkka%C5%A1a%20ahte%20%C2%ABl%C3%A1htte%20dego%20gumpe%C2%BB.%20%0A%0AD%C3%A1l%20leat%20ollu%20riikkat%20r%C3%A1f%C3%A1iduhtt%C3%A1n%20geatkki.%20Geah%C4%8Da%20Bernkonven%C5%A1uvnna%20%C2%ABListtu%20II%C2%BB%2C%20mii%20%C3%A1sahuvvui%201982%3As.%20Ruo%C5%A7as%20r%C3%A1f%C3%A1idahttojuvvui%20geatki1968%3As%20ja%20Lulli-Norggas%201973%3As.%20Eisev%C3%A1lddit%20leat%20mearridan%20unnimus%20logu%20boraspiriid%2C%20mat%20galget%20leat%20Norggas.%20Geatki%20d%C3%A1fus%20lea%20mihttomearri%20ahte%20galget%20leat%2039%20%C4%8Divgama%20jagis.%202010%3As%20leat%20dihtosis%2066%20%C4%8Divgi%20ciiko%2C%20muhto%20ma%C5%8B%C5%8Bel%20bivddu%20gos%20godde%20ciiko%20ja%20%C4%8Divggaid%2012%20biejus%2C%20de%20ledje%202010%20gi%C4%91a%2054%20biejus%20%C4%8Divggat" rel="nofollow">https:&#x2F;&#x2F;fanyi.baidu.com&#x2F;#sme&#x2F;en&#x2F;Geatki%20(Gulo%20gulo)%20lea...</a>.<p>(but at least &quot;Berners Convention&quot; is slightly close to the input …)<p>It&#x27;s amazing how low a bar big companies have for publishing ML models that just completely fail in the most absurd and spectacular ways, as if GM said &quot;hey marketing department it looks like people want electric cars now, that seems easy just strap a battery on a chicken and hammer some wheels onto it and call it a day&quot;
评论 #24620468 未加载
jacobwilliamroy超过 4 年前
I think it&#x27;s time to give up on machine translation. Let people do it. Maybe try and teach the computers how to do something we can&#x27;t already do ourselves.