www-legal-discuss mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Joern Kottmann <kottm...@gmail.com>
Subject Re: Training models for OpenNLP on the OntoNotes corpus
Date Fri, 17 Feb 2017 10:24:35 GMT
Hello all,

they replied to me and said the main issue is that their data (or models
trained on it) cannot be licensed under any agreements other than their
own. So this is the case for their research-only and commercial license.

Therefore training on LDC data (even if a member with the commercial
license would do it) and releasing the model under AL 2.0 (or any other
Open Source license) is not allowed.
On the other hand they seem to tolerate that Open Source projects are doing
that, when you google for models trained on their data you can find many
examples.

We will have to look for new sources of data to train our models on.

Thanks to everyone for helping with this issue.

Jörn


On Fri, Feb 17, 2017 at 9:06 AM, Peter Kluegl <pkluegl@gmail.com> wrote:

> Hi Joern,
>
>
> can you share the answer if you get one? I'd really appreciate it :-)
>
>
> Best,
>
>
> Peter
>
> Am 09.02.2017 um 15:56 schrieb Joern Kottmann:
>
> Hello,
>
> right, I agree with you, let me ask them.
>
> Thanks,
> Jörn
>
> On Wed, Feb 8, 2017 at 7:07 AM, Henri Yandell <bayard@apache.org> wrote:
>
>> The license says:
>>
>>     "In the event that User's use of the LDC Databases results in the
>> development of a commercial product, User must join...pay fees...".
>>
>> While I don't think LDC have necessarily considered Apache's use of their
>> product, and the license text doesn't appear to be considering a situation
>> where the two User definitions are different individuals (ie: Apache the
>> first, our users the second); I don't think it's clear that LDC are in
>> favour of our using their product and you should contact them to get
>> clarification that we can use their product to develop an Apache 2.0
>> licensed product which may subsequently be used in our user's commercial
>> products.
>>
>> Hen
>>
>> On Tue, Feb 7, 2017 at 7:01 AM, Joern Kottmann <kottmann@gmail.com>
>> wrote:
>>
>>> Thanks for your answer!
>>>
>>> We would not distribute the content itself in any way. The training
>>> process will reduce the input-copyright protected material into n-grams
>>> (which will have at most a length of 2). That work should not be
>>> copyright-protect able by the original copyright holder since we don't take
>>> anything out that is long enough to be able to be copyright protected.
>>>
>>> There was a case in the EU that might be relevant for this:
>>> https://en.wikipedia.org/wiki/Infopaq_International_A/S_v_Da
>>> nske_Dagblades_Forening
>>>
>>> Jörn
>>>
>>> On Mon, Feb 6, 2017 at 5:35 PM, Henri Yandell <bayard@apache.org> wrote:
>>>
>>>> I don't believe this acceptable.
>>>>
>>>> It's a non-commercial license that would restrict the uses of the
>>>> subsequent Apache product.
>>>>
>>>> Note that the license would also need signing (i.e. it's not something
>>>> we can use off the shelf).
>>>>
>>>> One approach would be to contact LDC to let them know our interest in
>>>> using, but make sure they understand that the output would be going into
a
>>>> product under the Apache 2.0 license and that they understand our concern.
>>>>
>>>> Hen
>>>>
>>>> On Fri, Feb 3, 2017 at 2:51 AM, Joern Kottmann <joern@apache.org>
>>>> wrote:
>>>>
>>>>> Hello all,
>>>>>
>>>>> the Apache OpenNLP library is a machine learning based toolkit for the
>>>>> processing of natural language text.It supports the most common NLP tasks,
>>>>> such as tokenization, sentence segmentation, part-of-speech tagging,
named
>>>>> entity extraction, chunking and parsing.
>>>>>
>>>>> Many of the competing solutions offer pre-trained models on various
>>>>> data sources to their users. We came to the conclusion that we have to
do
>>>>> the same to stay relevant.
>>>>>
>>>>> These corpora we would like to train on usually are copyright
>>>>> protected or have a license which restrict the use.
>>>>>
>>>>> I would like to know what the opinion here on legal-discuss is to
>>>>> train models based on the OntoNotes corpus [1]. Their license can be
found
>>>>> here [2].
>>>>>
>>>>> The training process does the following with the corpus as input:
>>>>>
>>>>> - Generates string based features (e.g. about word shape, n-grams,
>>>>> various combinations, etc.), those features to not contain longer parts
of
>>>>> the corpus text
>>>>>
>>>>> - Computes weights for those features based on the corpus
>>>>>
>>>>> The features and weights are stored together in what we call a model
>>>>> and this model we wish to distribute under AL 2.0 at Apache OpenNLP.
>>>>>
>>>>> Would it be ok to do that? Are there any concerns?
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Jörn
>>>>>
>>>>>
>>>>> [1] https://catalog.ldc.upenn.edu/LDC2013T19
>>>>>
>>>>> [2] https://catalog.ldc.upenn.edu/license/ldc-non-members-agreem
>>>>> ent.pdf
>>>>>
>>>>
>>>>
>>>
>>
>
>

Mime
View raw message