Michael Roth
@microth.bsky.social
92 followers 82 following 23 posts
Professor for Natural Language Processing (@utn_nuremberg), CoNLL co-chair 2025, organizer of LSDSem & UnImplicit workshops, expert in misunderstandings.
Posts Media Videos Starter Packs
microth.bsky.social
Want to see this guy in action again? Come to the Gender Bias in NLP (genderbiasnlp.bsky.social) poster session tomorrow, Aug 1, at 2pm! Joint presentation with PhD student Nicola Fanton, who is looking for a postdoc position soon. Come say hi!
Reposted by Michael Roth
conll-conf.bsky.social
Excited to see everyone tomorrow in Hall M1! 🎉 We kick off at 09:20 with the Opening. Don’t miss the 17:20 Discussion on:
👉 Handling growing submissions
👉Adapting conference formats
👉 Strengthening our community
Join us to shape the future! #CoNLL2025
microth.bsky.social
Finally on my way to #acl2025nlp as well, where I will co-chair this year's @conll-conf.bsky.social . My team will also share results on clarifying discourse relations (ACL Findings), classifying Spanish dialects (ACL-SRW) and a teaser for our 2026 WSD task at the #SemEval workshop. Come say hi!
microth.bsky.social
Our program is finally ready! =) #nlproc
conll-conf.bsky.social
🚨 CoNLL 2025 attendees!
🗓️ The tentative program schedule is now live on the website!
🎤 Don’t miss Raquel Fernández’s talk:
“The Interplay between Gestures and Speech in Face-to-Face Communication”
Questions? Let us know!
👇 Poster authors
#CoNLL2025
Reposted by Michael Roth
aachimova.bsky.social
More Ph.D. positions available in several projects related to the #SFB Common ground! Please see the list here uni-tuebingen.de/de/283575
Freie Stellen
uni-tuebingen.de
Reposted by Michael Roth
andreaskipf.bsky.social
The Data Systems Lab is seeking a motivated PhD candidate to join our team and work on foundation models for data compression.
microth.bsky.social
Is "citations per faculty" really a good metric for university rankings? Why don't we do "burnouts per publication" instead? Also, please don't cite me on this... :P
microth.bsky.social
Correct counts:
1700 accepted Main conference papers
1395 accepted Findings papers
108 accepted Industry papers
64 accepted demos
microth.bsky.social
According to ChatGPT, we will have a very small ACL this year.
chatgpt.com/share/6842a4...
Reposted by Michael Roth
barbaraplank.bsky.social
Do you speak Bavarian? Then please consider helping in a thesis research project👇
verenablaschke.bsky.social
Dei Boarisch heard ned bei "Servus" und "Pfiade" auf? Dann suach ma genau Di!
Wir suachan Bairischsprecher:innen, de a kurze Umfrage über KI-generierds Boarisch für a Masterarbeit beantwortn mechadn.
Mid jeder Teilnahme bring ma den boarischn Dialekt a Stickal weida in de digitale Weyd!
verenablaschke.bsky.social
Bavarian dialect speakers needed! Our MSc student Miriam wants to find out 1. how good/bad LLM-generated "Bavarian" is, and 2. whether dialect speakers agree with each other on this. The survey takes <5 min: survey.ifkw.lmu.de/dialquali25/ Thank you for sharing/participating!
microth.bsky.social
If you want to spare your "hard-working students" unnecessary rejections, you should not let them submit to venues that do not fit the topic of their work. This is your responsibility as a supervisor, not the fault of the venue's PCs...
microth.bsky.social
About 170 rejects and 40 accepts later, I just need to survive one more round of messages from disappointed authors. By the way, why do authors of accepted papers never write responses to decision notifications? #nlproc #conll2025 @conll-conf.bsky.social
microth.bsky.social
Two days after the meta-review deadline, 1 meta-review missing. Not so bad, eh? #conll2025 #nlproc @conll-conf.bsky.social
microth.bsky.social
One week later, we're still 8 reviews short, but we will make it. Many thanks to our fantastic area chairs and dozens of emergency reviewers! Also: good to hear that @emnlpmeeting.bsky.social is barring irresponsible reviewers from committing their work. Sadly right and necessary. #nlproc #conll2025
microth.bsky.social
Two days after the review deadline, 120 reviews are missing. Sent invitations to 186 emergency reviewers. About 20 have accepted so far. Organizing a conference is fun. Highly recommended. #conll2025 @conll-conf.bsky.social
microth.bsky.social
Of course, you can call your paper "The Leaderboard Illusion", but that's a bit misleading if it's only about problems with one particular leaderboard (Chatbot Arena). #arxiv
microth.bsky.social
Two days after the review deadline, 120 reviews are missing. Sent invitations to 186 emergency reviewers. About 20 have accepted so far. Organizing a conference is fun. Highly recommended. #conll2025 @conll-conf.bsky.social
microth.bsky.social
Lots of #nlp / #nlproc work these days is like: we generate synthetic data with GPT-4, fine-tune BERT, and then show that ourBERT performs better than BERT. But what does this really show? Is there anything meaningful that I am missing?
microth.bsky.social
Oh, a rejection email regarding a professor position for which I applied in 2023. I almost forgot about that one. #germanacademia #academictwitter
microth.bsky.social
There are people who, without being asked, suggest themselves as invited speakers for a conference??? 🤯
microth.bsky.social
It's really interesting to see what authors can get wrong when submitting papers: incorrect page sizes, wrong fonts, exceeded page limits, non-anonymized acknowledgements, entirely different paper templates, ...
microth.bsky.social
🚀 Attention #nlproc community! Got a standout *CL paper that has been flying under the radar? Our new non-archival submission track at #conll2025 will give you an extra spotlight. Let your research shine and submit your published paper here: docs.google.com/forms/d/e/1F... #NLP #ACL2025 #ARR
CoNLL 2025: non-archival submissions
Please see instructions for non-archival submissions on the CoNLL 2025 website.
docs.google.com
microth.bsky.social
We found two great invited speakers for this year's CoNLL!
conll-conf.bsky.social
🚨 Guest Speaker Alert! 🚨

We’re thrilled to announce that #CoNLL2025 will feature: 🥁

Raquel Fernández (University of Amsterdam)
&
Jean-Rémi King (@jeanremiking.bsky.social, CNRS / Meta AI)!
🎤✨

Check out their awesome work!👇
microth.bsky.social
Fun fact: In 2012, I took a truly inspiring class on Coursera called Model Thinking. Almost 13 years later, I am a professor myself and feel honored that @scottepage.bsky.social now follows me on Bluesky! His course is still available—check it out: www.coursera.org/learn/model-...