Machine translation is an area of Natural Language Processing concerned with automatically translating texts between different human languages.After a very optimistic start in the 1940s, Machine Translation (or MT as it is commonly abbreviated), hit a low with the publication of the U.S. ALPAC report (Automatic Language Processing Advisory Commitee) in the mid-60s. This report said that MT would never be as accurate or as cheap as translation by humans and so no more money should be wasted on it. After this, there was indeed very little research money spent on MT for quite a while, and a few of those who had been researching into MT instead started up companies and tried to pursue their interests there.
By the mid-70s, however, it was becoming clear that the advice of the ALPAC report was not going to remain accurate. Vast increases both in technology--e.g., the speed of computers and the quantity of memory available--and in our understanding of how language works, mostly coming from results in linguistics over the past 40 years, was beginning to make MT seem like a real possibility. The main trick is to find out where MT fits: it is not appropriate for all translation tasks, no one suggests MT for poetry for example; but there are places and niches where MT can be usefully employed.
A couple of introductory texts on MT can be found in the following:
There are a couple of introductory books that are quite useful, for example:
Some of the introductory overheads used in the course can be seen here.
A collection of MT systems--some very simple, some more complex--are given on the 'MT hands on' page for this course. For the lab session you should go to this page and follow the directions there.