Join the conversation

Join the community of Machine Learners and AI enthusiasts.

Sign Up
nicolay-rΒ 
posted an update Oct 6
Post
643
πŸ“’ This year I made decent amout of experiments on LLM reasoning capabilities in author opinion extraction.
However, they did not go further with:
↗️ annoation of other sources of opinion causes: entities, out-of-context object (None).
πŸ“ evaluation of factual statements that support the extracted sentiment.

To address these limitations, so far we launch πŸš€ RuOpinionNE-2024 competition on the Codalab platform:
πŸ“Š https://codalab.lisn.upsaclay.fr/competitions/20244

The competition is aimed at extraction of opinion tuples (see attached images) from texts written in Russian.
It proceeds the past RuSentNE-2023 codalab competition findings:
πŸ”Ž Past year competition: https://www.dialog-21.ru/media/5896/golubevaplusetal118.pdf
πŸ”Ž LLM reasoning 🧠: https://arxiv.org/abs/2404.12342

For those who interested to adopt Generative AI, the complete information about competition is below:
πŸ“Š RuOpinionNE-2024: https://codalab.lisn.upsaclay.fr/competitions/20244
πŸ—’ Task description: https://codalab.lisn.upsaclay.fr/competitions/20244#learn_the_details-overview
πŸ”” To follow updates: https://t.me/RuOpinionNE2024
⏰ Stages Deadlines (might be extended)
πŸ“¦ Submission details (bottom of the competition page)

πŸ™‹ For questions you can contact @nicolay-r : https://nicolay-r.github.io/
πŸ§ͺ Most recent findings on LLM application: https://github.com/nicolay-r/RuSentNE-LLM-Benchmark
In this post