Progress in AI continues to outpace benchmarks.
Check out this new plot, inspired by @DynabenchAI, that shows just how quickly it's happening.
Read more about it here: https://t.co/MwmfOdUy0B
I was puzzled for a while as to why we need RL for LM training, rather than just using supervised instruct tuning. I now have a convincing argument, which is also reflected in a recent talk by @johnschulman2 . I summarize it in this post:
https://t.co/DQD1wgyjg3
GPT-4 and its ilk are awesome for rapid prototyping and one-offs, but at the end of the day, enterprises will deploy far smaller distilled models in production. Here's my contrarian take -
Congrats to @ArthurBrack and team for the Best Student Paper #JCDL2022
"Cross-Domain Multi-Task Learning for Sequential Sentence Classification in Research Papers"
Very happy to announce that we have won the "best student paper award" at the @jcdl2022 . Yeah!!! I am so excited and thankful about it! Big thanks go to the JCDL 2022 and my co-authors @jeary_talking, Pascal Buschermöhle and @RalphEwerth. https://t.co/hxlOOpzrmL
#jcdl2022#orkg
We reached the 10,000 paper mark!🥳A big ‘thank you’ to our over 700 users who made the effort to describe their research semantically and feed the ORKG with content! Let’s keep up the good work and build a new scholarly communication platform together. Next goal: 20,000 papers🚀
Last call for **Late-breaking results, Preliminary work, Datasets, and Demonstrations** papers @jcdl2022#JCDL2022
The deadline is coming close: March 21 (AoE)!
https://t.co/CrIQMmOQ9b
Chairs: @feedkoko@nicvtwoz@dwaipayanroy
The list of accepted papers / research track @jcdl2022 is online: https://t.co/OCMqyOvjMg
Congrats to all authors! #jcdl2022
--> The late breaking, dataset and demo track (short papers) is open till **March 21** (AoE)
https://t.co/99h3ZvKN55 Keep up submitting!
Happy to announce that our research paper "Cross-Domain Multi-Task Learning for Sequential Sentence Classification in Research Papers" co-authored with @jeary_talking, Pascal Buschermöhle, and @RalphEwerth for the @orkg_org has been accepted for publication at @jcdl2022
With Russian colleagues, we created a document refuting Kremlin propaganda myths.
Russian: https://t.co/q8HozlwS02
English (autotranslated): https://t.co/ZF1KT1iacC
Plan to share this widely with contacts in Russia - any improvement suggestions?
#StandWithUkraine️
Just had a long discussion with Russian friends on IM - although opposition-minded they are completely unaware of the extent this terrible war has on Ukraine, Russia and the world. If you have friends in Russia, please confront them with the reality (thread). #StandWithUkraine
Here's the December 2021 release of draft chapters for Speech and Language Processing, just in time for Jim and I to wish you all a Happy New Year! Enjoy! https://t.co/pjfZsYfJ5S
Happy to announce the release of the new @Open_access public data dump of the OpenAIRE Research Graph (https://t.co/CT2Fa1XLxj) in @ZENODO_ORG (https://t.co/ZBXtmy99zM)! All known & trusted metadata sources interlinked and deduplicated! Happy New Year from the @OpenAIRE_eu team!
@MFordFuture Huge systems like GPT-3 are astonishing.
They're just the start for what @percyliang and colleagues call "Foundation Models" reflecting their broad applicability.
Here's an overview of key issues including a chapter on economic implications:
https://t.co/KRtKXIy0Ff
@DigEconLab
The three studies explore: Gopher - a SOTA 280B parameter transformer, ethical and social risks, & a new retrieval architecture with better training efficiency.
1: https://t.co/WDUeFd5DiF
2: https://t.co/cZcWHCg128
3: https://t.co/h9fdMP6C5W (more https://t.co/4QiVDqntTS) 2/