[Corpora-List] Fw: Competition Launch: Contradictory, My Dear Watson!

Eric Atwell E.S.Atwell at leeds.ac.uk
Fri Aug 7 09:49:50 CEST 2020


This new Kaggle contest is directly related to Corpus Linguistics: "Given pairs of sentences (a premise and a hypothesis), you will predict whether they are related. For an added challenge, the train and test set include text in fifteen different languages!"

Please take a look at this challenge and consider taking this on, as an interesting exercise. There will be a Kaggle community of researchers working on this, and a discussion forum to share ideas. At minimum, you may learn some new techniques which you can then apply to your own corpus linguistics research. I am also thinking about using this data-set and task as an exercise in online teaching next year.

eric

Eric Atwell, Professor of Artificial Intelligence for Language

PhD tutor; online MSc AI programme leader

School of Computing, Uni of LEEDS, LS2 9JT, UK

http://www.comp.leeds.ac.uk/eric https://www.edubots.eu/

http://lists.leeds.ac.uk/mailman/listinfo/arabic-l http://corpus.quran.com

________________________________ From: Kaggle <noreply at kaggle.com> Sent: 07 August 2020 00:53 To: eric at comp.leeds.ac.uk <eric at comp.leeds.ac.uk> Subject: Competition Launch: Contradictory, My Dear Watson!

[https://storage.googleapis.com/kaggle-media/email-templates/kaggle-header.png] <https://www.google.com/appserve/mkt/p/AD-FnEzRkVg9kXZ9NYKKwWb63dRn5Yp-MeLZTFKrCody61AaSNnKLXXyh7P0has60DoxaX_5UsIF8tiXVTQ>

Hi ericatwell,

How quickly can you distinguish whether two sentences are related? Our brains do this surprisingly well. It’s a more difficult problem for a computer, which is the focus of a new Getting Started Kaggle Competition. We’re even providing a starter notebook<https://www.google.com/appserve/mkt/p/AD-FnEz5-8cCuHG_Mud2ULyrrnnNLg9KV5zyj3AKloFGSRunpb9wP65t1mEkOanyN-lu-M65d1fIVOfSqQTnPqr_V92mVN8NCozosEdPN68UGdSYr-AmRfXfpcVp> to try your hand at this problem using the power of Tensor Processing Units (TPUs).

You will use Natural Language Inferencing (NLI), a popular Natural Language Processing (NLP) approach. Given pairs of sentences (a premise and a hypothesis), you will predict whether they are related. For an added challenge, the train and test set include text in fifteen different languages!

Join the Competition <https://www.google.com/appserve/mkt/p/AD-FnEw8A1yOlXRWjYjtaksVEsbMD47cuK2mjrF42-r5REtH38lQ0EuWl8VRb6Nf8tCwn74GNRPIvCsy7N0RQBX6BTzKi6CBthYXAZL_jeCYr_Uf1ib34VxD2Bw>

This Getting Started competition will be ongoing and live for anyone to participate at any time. But TPU Star Awards<https://www.google.com/appserve/mkt/p/AD-FnEyyMtsaKZ4QHrl0B3SDdT3gNTjU3sGO1kdzEEUkQWB-23GVKK_GzntXEk1byIoTgHuNLzraL4Mw0CGLFiKKriBAQ0QSdKKGYYVP93Y6RptHkLOPWOLSxx0CkBQZTyZzco8qqmPoVOzP> for extra TPU quota are only open for self-nomination until September 30, 2020! This is a great opportunity to flex your NLP muscles and solve an exciting problem!

Thanks and good luck, Phil Culliton Kaggle Data Scientist

Kaggle, Inc 1600 Amphitheatre Pkwy Mountain View, CA 94043

This email was sent to eric at comp.leeds.ac.uk because you indicated that you'd like to receive news and updates about Kaggle. If you don't want to receive these emails in the future, please unsubscribe here<https://www.google.com/appserve/mkt/p/AD-FnEyf5biZT9ZJnIxBfCsOEJnWG43U6UOQ0f_CGPtK86vjT9Km0-2v8zQre9gXthTf5XT0fdEPE_Ogx1sY0NlFBP04-O3NPU-SmeGeai1n8sB0MHzWc2JqStiimulQnSWp8Ub4JofUDe7hV26XKGZDZCDwK1LDniFOkWRmD99FpMGSTNF9GZnH>. You can also change your preferences on your account's profile page by logging in at kaggle.com<https://www.google.com/appserve/mkt/p/AD-FnEzRkVg9kXZ9NYKKwWb63dRn5Yp-MeLZTFKrCody61AaSNnKLXXyh7P0has60DoxaX_5UsIF8tiXVTQ>.

-------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/html Size: 12268 bytes Desc: not available URL: <https://mailman.uib.no/public/corpora/attachments/20200807/0cb2d3f4/attachment.txt>



More information about the Corpora mailing list