Introduction to Statistical Machine Translation
michaelnielsen.org
michaelnielsen.org
This really struck me, as someone who both teaches and studies language. If it works better for computers, I wonder how much better it would work for people. If there's anything I've noticed in my time in Japan, it's that the Japanese approach seems to be to nail down every single word with a single Japanese meaning and stick to that meaning all the time, which leads to a lot of very Japanese-sounding English.
Ouch.
Building MT systems without knowing foreign languages is a bit like deaf people building a speech recognizer. No offence meant to anyone, but it works much better when you know what you're doing.
IIRC the 'candide' group was (not intentionally) composed of scientists with no knowledge of both english and french.. http://www.cs.cmu.edu/~aberger/mt.html
Contrary to your point, early statistical machine translation only works well for relatively close language pairs, like English-French or English-Spanish. It totally fails for more distant languages such as Chinese, Arabic or even German, which is why you have so many Chinese-speaking people (including English-Chinese bilinguals) in machine translation these days.
Parallel corpora are full of ad-hoc, hard-to-model, human judgements (from people called "translators"). The advantage is that the translators don't come up to you to criticize your translation model; however doing error analysis for an MT system (i.e., the key to actually improving things and not producing garbage) requires at least minimal knowledge of the source language and relatively good knowledge of the target language.