Hello John,
In the domain of natural language generation, there has been some work around generating text from Wikidata triples (mostly for Wikipedia):
In under-resourced languages, domain-independent: https://link.springer.com/chapter/10.1007/978-3-319-93417-4_21
In under-resourced languages, domain-independent: http://aclweb.org/anthology/N18-2101
In English, for the biography domain (Wikidata and DBpedia): https://www.sciencedirect.com/science/article/pii/S1570826818300313
In English, for the biography domain: https://arxiv.org/abs/1702.06235
Since we worked in this field, please reach out if you have questions regarding those publications. Best, Lucie
On Wed, 23 Jan 2019 at 13:58, john cummings mrjohncummings@gmail.com wrote:
Hi all
I'm putting together an overview of Wikimedia for a conference submission and their theme is AI. Does anyone have any examples of use of Wikimedia by AI projects?
Thanks
John _______________________________________________ Wiki-research-l mailing list Wiki-research-l@lists.wikimedia.org
https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.wik...
Coincidentally, Louise Matsakis also just wrote a piece for Wired about the Google donations in which she includes a bunch of pointers to how Wikimedia content is used by AI systems: https://www.wired.com/story/google-wikipedia-machine-learning-glow-languages...
I would also point you towards a few high-profile instances: Major dataset used for training AI question and answer systems: https://rajpurkar.github.io/SQuAD-explorer/ Entity linking from OpenAI: https://blog.openai.com/discovering-types-for-entity-disambiguation/ Scanning through dataset search might help as well: https://toolbox.google.com/datasetsearch/search?query=wikipedia
Best, Isaac
On Wed, Jan 23, 2019 at 10:13 AM Lucie-Aimée Kaffee kaffee@soton.ac.uk wrote:
Hello John,
In the domain of natural language generation, there has been some work around generating text from Wikidata triples (mostly for Wikipedia):
In under-resourced languages, domain-independent: https://link.springer.com/chapter/10.1007/978-3-319-93417-4_21
In under-resourced languages, domain-independent: http://aclweb.org/anthology/N18-2101
In English, for the biography domain (Wikidata and DBpedia): https://www.sciencedirect.com/science/article/pii/S1570826818300313
In English, for the biography domain: https://arxiv.org/abs/1702.06235
Since we worked in this field, please reach out if you have questions regarding those publications. Best, Lucie
On Wed, 23 Jan 2019 at 13:58, john cummings mrjohncummings@gmail.com wrote:
Hi all
I'm putting together an overview of Wikimedia for a conference submission and their theme is AI. Does anyone have any examples of use of Wikimedia
by
AI projects?
Thanks
John _______________________________________________ Wiki-research-l mailing list Wiki-research-l@lists.wikimedia.org
https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.wik...
-- Lucie-Aimée Kaffee Web and Internet Science Group School of Electronics and Computer Science University of Southampton _______________________________________________ Wiki-research-l mailing list Wiki-research-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Thanks very much :)
On Wed, 23 Jan 2019 at 16:30, Isaac Johnson isaac@wikimedia.org wrote:
Coincidentally, Louise Matsakis also just wrote a piece for Wired about the Google donations in which she includes a bunch of pointers to how Wikimedia content is used by AI systems:
https://www.wired.com/story/google-wikipedia-machine-learning-glow-languages...
I would also point you towards a few high-profile instances: Major dataset used for training AI question and answer systems: https://rajpurkar.github.io/SQuAD-explorer/ Entity linking from OpenAI: https://blog.openai.com/discovering-types-for-entity-disambiguation/ Scanning through dataset search might help as well: https://toolbox.google.com/datasetsearch/search?query=wikipedia
Best, Isaac
On Wed, Jan 23, 2019 at 10:13 AM Lucie-Aimée Kaffee kaffee@soton.ac.uk wrote:
Hello John,
In the domain of natural language generation, there has been some work around generating text from Wikidata triples (mostly for Wikipedia):
In under-resourced languages, domain-independent: https://link.springer.com/chapter/10.1007/978-3-319-93417-4_21
In under-resourced languages, domain-independent: http://aclweb.org/anthology/N18-2101
In English, for the biography domain (Wikidata and DBpedia): https://www.sciencedirect.com/science/article/pii/S1570826818300313
In English, for the biography domain: https://arxiv.org/abs/1702.06235
Since we worked in this field, please reach out if you have questions regarding those publications. Best, Lucie
On Wed, 23 Jan 2019 at 13:58, john cummings mrjohncummings@gmail.com wrote:
Hi all
I'm putting together an overview of Wikimedia for a conference
submission
and their theme is AI. Does anyone have any examples of use of
Wikimedia
by
AI projects?
Thanks
John _______________________________________________ Wiki-research-l mailing list Wiki-research-l@lists.wikimedia.org
https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.wik...
-- Lucie-Aimée Kaffee Web and Internet Science Group School of Electronics and Computer Science University of Southampton _______________________________________________ Wiki-research-l mailing list Wiki-research-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
-- Isaac Johnson -- Research Scientist -- Wikimedia Foundation _______________________________________________ Wiki-research-l mailing list Wiki-research-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
wiki-research-l@lists.wikimedia.org