Google today open sourced the most accurate parser (they claim!) - Syntaxnet.
Post: https://research.googleblog.com/2016/05/announcing-syntaxnet-worlds-most.html?m=1
Facebook open sourced their version of Neural Translation System using CNNs.
Post: https://code.facebook.com/posts/1978007565818999/a-novel-approach-to-neural-machine-translation/.
GitHub: https://github.com/facebookresearch/fairseq
While traditionally translation tasks were done using RNNs, they claim the novel CNN approach beat the other approaches by 9x the speed and are more accurate as well.