[Moses-support] interface for subjective evaluation

Hi all, I would like to set up a web-based subjective evaluation of some SMT systems. I was wondering if there exist an open source interface (in php, javascript...) to do this without using Mechanical Turk or CrowdFlower. Possibly something simple that I can customize for error annotation Thanks

[Moses-support] bad word alignment in phrase decoder

y the IOWrapper to print out the alignment in the nbest... but ther might be others. Best, Arianna Bisazza ___ Moses-support mailing list Moses-support@mit.edu http://mailman.mit.edu/mailman/listinfo/moses-support

Re: [Moses-support] Is Moore & Quirk (2007a) implemented in Moses?

expected improvement by using a DL slightly higher than my baseline. Best, Arianna Bisazza > On Tue, Mar 6, 2012 at 10:59 AM, Lane Schwartz wrote: >> Moore & Quirk (2007), presented at MT Summit XI, describe a mechanism >> for incorporating a distortion penalty estimate in the fu

[Moses-support] Compilation fails "--without-kenlm"

Hi all, I've noticed a problem in the version of Moses just downloaded from Github. If I configure with the option "--without-kenlm", compilation fails apparently because the directory "util" cannot be found. When I included kenlm, compilation succeeded instead. Che

[Moses-support] Verbose 3 logging heavily increased

s the change's comment: Added additional level 3 logging to help verify that Syntactic LM is being loaded and used Could this be made conditional on the presence of a Syntactic LM at least? Cheers, Arianna Bisazza FBK-irst ___ Moses-support ma

[Moses-support] Strange behaviour of AlignmentInfo::NonTermIndexMap

phrase-based decoding context. Is this a bug? If not, can this object be used at all outside the chart-decoding context? Just for clarification, what I'd like to do is simply enumerating the (aligned) source positions in target order - in a text input context. Thanks, Arianna Bisazza FBK