Plenty of resources, but you have to pick your source and target
language first (unless you want true statistical translation and have
a true Google-size corpus to train it on).
Once you know your languages, then you can talk about resources.
Machine Translation is one of the more complex part of NLP/CL, so you
need to define your interest quite narrowly. And also whether you are
looking for practical results ('good enough for use') or research
progress with validation scores, etc.
Regards,
Alex.
Personal blog:
http://blog.outerthoughts.com/
Research group:
http://www.clt.mq.edu.au/Research/
- I think age is a very high price to pay for maturity (Tom Stoppard)