Papers of the day   All papers

KILT: a Benchmark for Knowledge Intensive Language Tasks

Comments

Fabio Petroni: I'm excited to present KILT, a benchmark for Knowledge Intensive Language Tasks. 11 datasets, 5 task families, a unified knowledge source and interface -> https://arxiv.org/abs/2009.02252. Collaboration b/w @facebookai @ucl_nlp @labo_Loria @cambridgenlp @huggingface @AmsterdamNLP 🚀 [1/6] https://t.co/VEjUDf7now

3 replies, 418 likes


Patrick Lewis: We’re releasing KILT today, a set of tools and data for knowledge intensive NLP. It will make it easier to develop and evaluate systems that need to retrieve and process knowledge precisely. We’re really excited about this effort! Big kudos to the KILT mastermind @Fabio_Petroni

0 replies, 82 likes


Antoine Bordes: KILT or how to foster progress in Knowledge-based NLP by combining 5 types of tasks, 11 datasets with the same knowledge source (6 millions Wikipedia articles)!

0 replies, 32 likes


James Thorne: Was a pleasure to work with a great team lead by @Fabio_Petroni at @facebookai building KILT - a benchmark for Knowledge Intensive Language Tasks. What an awesome way to link so many NLP tasks against a common @Wikipedia knowledge source

0 replies, 27 likes


Yacine Jernite: I am so excited for this work by @Fabio_Petroni @olapiktus and many other cool folks, it will be an invaluable tool to develop models that reliably use knowledge from large text bases. You can even get the data in just a couple LOC from🤗nlp, check it out! https://huggingface.co/datasets?search=kilt

0 replies, 23 likes


WikiResearch: A benchmark for knowledge-intensive language tasks (KILT), all grounded in the same snapshot of #Wikipedia.

0 replies, 21 likes


Tim Rocktäschel: Important step for evaluating and developing general models that learn to utilize human knowledge in large textual collections like Wikipedia 👇

0 replies, 18 likes


Ola Piktus: KILT - a benchmark for Knowledge Intensive Language Tasks goes live today 🎉 Congrats @Fabio_Petroni and the team!

0 replies, 16 likes


Nicola De Cao: KILT a new benchmark for knowledge intensive tasks! Super nice work by @Fabio_Petroni and others at @facebookai of unifying so many datasets under the same format and aligning them to the same knowledge source! Let's start tracking progress of retrievers and NLU system on KILT 😊

0 replies, 14 likes


AK: KILT: a Benchmark for Knowledge Intensive Language Tasks pdf: https://arxiv.org/pdf/2009.02252.pdf abs: https://arxiv.org/abs/2009.02252 https://t.co/ZCeXfFnPPJ

0 replies, 8 likes


HotComputerScience: Most popular computer science paper of the day: "KILT: a Benchmark for Knowledge Intensive Language Tasks" https://hotcomputerscience.com/paper/kilt-a-benchmark-for-knowledge-intensive-language-tasks https://twitter.com/Fabio_Petroni/status/1302970490265505793

0 replies, 7 likes


FEVERworkshop: We're excited to announce that the @FEVERworkshop dataset is part of the KILT benchmark with @facebookai

0 replies, 6 likes


Leo Boytsov: "To catalyze research on models that condition on specific information in large textual resources we present benchmark for knowledge-intensive language tasks (KILT): task are grounded in the same snapshot of Wiki, reducing engineering turnaround through the re-use of components"

0 replies, 4 likes


HubBucket: #HealthIT | Download the #KILT #Research Paper for FREE, on @arxiv #KILT is a Benchmark for Knowledge Intensive Language Tasks 🔵Natural Language Processing - #NLP/#NLProc 🟣#MachineLearning - #ML 🖥️https://arxiv.org/abs/2009.02252 | @FacebookAI @HubBucket @ProsumerSoft @HubMedX https://t.co/Qzj4nCnHZN

0 replies, 4 likes


Content

Found on Sep 07 2020 at https://arxiv.org/pdf/2009.02252.pdf

PDF content of a computer science paper: KILT: a Benchmark for Knowledge Intensive Language Tasks