On 9/2/11 10:05 AM, Nicolas Hernandez wrote:
Yes, Jörn.
Since we (LINA Lab.) managed to get the right to distrisbute the
models, we contacted the owner (actually her legal department) in
July, to ask her if it would be possible, with respect to the corpus
rights, to specify in the license that the trained models could be
distributed by anyone under AL2.
So far, no news. I will ask again soon.
Great, I think having this exception in their license would be the
best solution for everyone, I guess we then don't even need
all the papers.
Usually these statistical models only contain at best short n-grams
which cannot be used to re-construct the original text and don't
violate the copyright of the training material it was created from.
Jörn
