#PhD job in the Dept. of Language and Information Sciences at the University of Lausanne: my colleague Davide Picca has an open PhD position starting on October 1, 2025 in an SNSF-funded project focused on the computational analysis of Charles S. #Peirce’s manuscripts.
Deadline for application: May 19, 2025
Hello! My husband @trurl has been using Mastodon for a while, and I finally decided to make an account as well.
About me:
* Software engineer in #NLProc / #MachineLearning space
* #BikeCommuting (#ebike , to be exact)
* interested in local issues for #SomervilleMA and #Boston area
* I also enjoy #knitting as well as building and painting #gunpla figures
Look you may say that I can’t do interesting research because it’s not tied to some specific business milestone, but I’m going to still sneak it in anyway…
And if it’s esoteric enough that my slow side-project pace won’t be scooped by some other lab with more flexibility, so much the better!
#AcademicInIndustry #AIML #NLProc #IndustryResearcher
Moved all my stuff out of Dropbox (didn’t have much there), Google drive is next (but it’s a bit more messy and complicated).
I have a colleague who years back was concerned about keeping work stuff (eg paper drafts, grant proposals) on Google -we’re in #nlproc, kind of the same area as they are-, and I thought he was a bit paranoid. Now I think it’s probably best to keep our stuff closer to home instead of on US clouds. #academicChatter #europe #warOnScience
Final Call for Nominations: Inaugural Computational Linguistics Doctoral Dissertation Award to recognize and promote outstanding doctoral research in the field of computational linguistics and NLP. Deadline: Jan 31, 2025. Nominations Details: https://bit.ly/cl-dissertation-award #NLProc
I am spending Boxing Day working on my @everythingopen.au #EO2025 talk on "Token Wars" and doing a little light reading to brush up.
Your attention is all I'll need 
1/
How well can Transformers build World Models?
Large Language Models (#LLMs) are statistical in nature. By learning from enormous corpora, do they actually learn “world models” latent behind the texts? Can next-token prediction really learn the principles governing the environment it operates in?
My lab, Computational Linguistics at Manitoba, is seeking motivated PhD students for #AI and #NLProc research in computational humour, historical born-digital #corpora, and #Indigenous language technology: https://clam.cs.umanitoba.ca/open-positions
...and here's another announcement of a virtual event:
On October 24th, 15:00 (CET), Ekaterina Shutova (Associate Professor at the University of Amsterdam) will talk about cross-lingual information sharing in multilingual language models.
The session is hosted under the roof of AIDA, the artificial intelligence doctoral academy (an institution we've cooperated with in AI4MEDIA).
More info & (free) registration:
https://www.i-aida.org/events/cross-lingual-information-sharing-in-multilingual-language-models/
Tfw you’re just casually exploring some new approach and lib installs lead you to an active cyber security flag in nltk, because pickle. 
It seems that we're getting close to a universal translator, which will have significant impact on international media.
When #Google Translate started in 2006, it offered translations in only a few languages – in pretty poor quality. They refined the service, switched to neural networks in 2016, added more models, tweaked it again – and now support 243+ languages (mostly with decent results). Google's goal is to eventually implement 1000 (!) languages.
Day 3 of daily 10k #running in #NAACL2024 in Mexico City: we ran a bit slower than yesterday on our way to the park (~10’/mile), and today we ran up to the hill! Pace picked up on our way back — up to ~8’20”/mile. I took a tumble on a pothole, and practiced perfect rolling action on the ground. No biggie — finished my daily 7-mile quota.
From an email I received this morning from the company that has been providing AV services to ACL events with the subject "How Can you Help Advance Open Science ?" Folks, videos ARE NOT ANONYMOUS. I will definitely not be approving this and caution all #NLProc researchers who have received similar messages from doing so. /1
![Screencap of email:
Dear Emily M.
We are contacting you because you previously presented in conferences held in Underline. You can access your presentations by visiting your profile: [URL redacted]
Imagine that now, your presentation can keep giving – and help your fellow researchers, further advance the state of AI and linguistics – even while you are sleeping!
We have been contacted by researchers who would like to use samples of your talk to help train their data models.
Note: This is completely anonymous, with no identification of you or your talk to any outside party ever.
This will help advance the state of open science and provide foundational models for others to use.
Please click to indicate your acceptance:
--
I have highlighted the false statement about anonymity with a red line.](https://cdn.masto.host/daircommunitysocial/media_attachments/files/111/414/892/896/117/700/original/105901507944d2b6.png "Screencap of email:
Dear Emily M.
We are contacting you because you previously presented in conferences held in Underline. You can access your presentations by visiting your profile: [URL redacted]
Imagine that now, your presentation can keep giving – and help your fellow researchers, further advance the state of AI and linguistics – even while you are sleeping!
We have been contacted by researchers who would like to use samples of your talk to help train their data models.
Note: This is completely anonymous, with no identification of you or your talk to any outside party ever.
This will help advance the state of open science and provide foundational models for others to use.
Please click to indicate your acceptance:
--
I have highlighted the false statement about anonymity with a red line.")
Very excited to share a substantially updated version of our preprint “Language models show human-like content effects on reasoning tasks!” TL;DR: LMs and humans show strikingly similar patterns in how the content of a logic problem affects their answers. Thread: 1/10
#LanguageModels #lms #AI #cogsci #machinelearning #nlp #nlproc #cognitivescience
