Carsten Schnober

Natural language processing, machine learning, deep learning, AI.
Software engineering, architecture, design.

[CLIN33 @CLIN33conf]

A full house at the #clin33 keynote by @IGurevych at @UAntwerpen on NLP for fact checking.

[Carsten Schnober @CarSchno]

#huggingface #transformer #LLMs
https://t.co/zW1HvZ6sto

[Carsten Schnober @CarSchno]

Einstieg in #PyTorch
https://t.co/H4MZVMKzMZ

[Netherlands eScience Center @akademienl.social @eScienceCenter]

📢 The @esciencecenter released its RSE job profile & role description to bring recognition to #researchsoftware & RSEs! They can serve as resources for those looking to define & appropriately position RSEs w/in their org. #RecognitionRewards

👉https://t.co/a1QMFS9Sxe

Please RT

[Carsten Schnober @CarSchno]

Ich durfte auch etwas beisteuern

[iX @iXmagazin]

iX Special "Künstliche Intelligenz" ist da https://t.co/WkQZurD07Q #Datenschutz #DSGVO #Grafikchip #KünstlicheIntelligenz #MachineLearning #Recht

[Netherlands eScience Center @akademienl.social @eScienceCenter]

#LanguageModeling is trending, to a large extent because of #ChatGPT. But did you know language modeling has been with us for more than a century? And that it was born of the collaboration of a poet and a mathematician? Our engineer @carschno tells us more:https://t.co/IZGX7lFf1Q

[Alex Volkov @ AI Engineer @altryne]

Watching @karpathy presentation from today and taking twitter notes, come along for the ride:

If you're like only the practical tips, skip to #32

@karpathy starts with stages:
1 - Pre-training - months x thousands of GPUs
2, 3, 4 - Finetuning stages that take hours or days

1/

[Isabelle Augenstein @IAugenstein]

Thanks to @egrefen for his great #eacl2023 keynote!
Some take-aways:
- LLMs struggle to grasp pragmatics
- data quality and diversity are key to LLM performance
- open-ended RL can help with automatic data selection
#NLProc

[(((ل()(ل() 'yoav))))👾 @yoavgo]

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

[Aaron Benanav @abenanav]

Is ChatGPT coming for your job? Catch me on @techwontsaveus talking to @parismarx about the robot jobs-apocalypse that wasn't and what that might tell us about generative AI and the future of work. https://t.co/NlBbIaNHDG

[Carsten Schnober @CarSchno]

@KLM I've just booked a flight, but the server died during the process ("session expired"). I see the flight in my trips, along with a booking code, but I see no payment nor do I get an invoice.
Customer service is not reachable. How can I get in touch?

[Netherlands eScience Center @akademienl.social @eScienceCenter]

❗Researchers now have a way to manage software as one of the core outputs of their research. In an initiative led by us & @NWONieuws w/ >80 contributors around the 🌍, we're finally launching the Practical Guide to Software Management Plans!

Learn more:
https://t.co/mFWxmcsRyw

[Ben Carterette @BenCarterette]

As promised, here’s the open access #sigir2022 @SIGIRConf proceedings! Open access is available only via this page, not directly thru the Digital Library, so please share the link widely! https://t.co/LRdUf0ERg2

[Carsten Schnober @CarSchno]

ICLR conference keynote on AI interpretability.

[Been Kim @_beenkim]

I've uploaded this keynote video to: https://t.co/rqVRgiqTtC Thank you organizers for granting me to upload earlier. This talk + QnA will be also available in a few weeks at ICLR website: https://t.co/fIz3zHHlWl

[Sole Pera @DrCh0le]

Great to see @WizeNoze sharing their work on IR for young learners #ECIR2022

[Wizenoze @WizeNoze]

At ECIR 2022, the 44th European Conference on Information Retrieval, Carsten Schnober presented the Wizenoze "Semantic Search" project: https://t.co/Vy2O7ks2QZ. #edtech

[Negar Arabzadeh @NegarEmpr]

We got the best short paper award @ecir2022 🤩 I’m super greatful to be a part of this team!!! @amin_bigdelii @shirin_ssalehi_ @MortezaZihayat @ebrahim_bagheri
#ecir2022

[Daphne Ippolito @daphneipp]

We noticed that the recent arxiv paper "A Roadmap for Big Model" with 100 authors plagiarizes from many papers, including ours. Nicholas Carlini wrote a blog post about it: https://t.co/KEHGMu9yQD

[Nicola Ferro @frrncl]

Great attendance at #ecir2022 @ecir2022 both in-presence and remotely!