most probably, he has already discussed about all issues related to his research with his supervisors at the Technische Universität here in Berlin.
Let us not start again intimidating young researchers, I really want to ask you very kindly!
Would you, ted, know where to get access to the corpora required in the original mail?
With kind regards, Valia Kordoni
On Mon, June 21, 2021 14:08, Ted Pedersen wrote:
> Hi Daniel,
> I might suggest some caution in pursuing this data set. I think there
> are quite a few questions about it regarding possible terms of service
> violations (among others). You could encounter some difficulties
> either with your institution, program committees, or journal editors
> should you seek to defend or publish results based on this data, and
> it isn't clear if you'd be able to redistribute the data (so your
> results may not be reproducible).
> In any case, I would suggest discussing the use of this data fairly
> carefully with folks at your institution before proceeding too deeply
> with it.
> Dr. Ted Pedersen
> UEA-D Member : https://sites.google.com/ueaumd.org/ueaumd
> On Mon, Jun 21, 2021 at 1:08 AM Fernau, Daniel
> <daniel.fernau at campus.tu-berlin.de> wrote:
>> Dear Community,
>> I am looking for the myPersonality Dataset containing Big Five
>> personality scores of about 3.1 million users from David Stilwell and
>> Michal Kosinski for my master thesis.
>> AIMS AND SCOPE:
>> I am currently writing my master thesis at Technische Universitï¿½t
>> Berlin where we are investigating possibilities of adaptive
>> conversational agents. Next to emotion, we are planning to examine the
>> impact of personality adaption of conversational agents where we need to
>> assess the personality of the user from the interaction from text. To
>> recognize personality from text, I am planning to fine-tune a BERT
>> model, but datasets containing personality scores are scarce.
>> Nevertheless, many papers on related topics are based on the
>> myPersonality dataset from David Stilwell and Michal Kosinski
>> Unfortunately, they stopped sharing their dataset in 2018 due to a lack
>> of capacity for maintaining and responding to the inquiries.
>> Therefore, I would like to train my model with a strong dataset to
>> achieve competitive results.
>> If you have maintained a copy of the corpus or can send me a pointer
>> where I can obtain the corpus, I would be very grateful!
>> Thank you in advance!
>> UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
>> Corpora mailing list
>> Corpora at uib.no
> UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
> Corpora mailing list
> Corpora at uib.no