Ece Takmaz (@ecekt2) header
Ece Takmaz (@ecekt2) profile picture

Interim Professor for AI and Computational Linguistics @LMU_Muenchen Previously Postdoc at @UniUtrecht, PhD at @UvA_Amsterdam

Munich, Germany
Joined August 2017
Photo by Ece Takmaz (@ecekt2)Photo by Ece Takmaz (@ecekt2)Photo by Ece Takmaz (@ecekt2)Photo by Ece Takmaz (@ecekt2)Photo by Ece Takmaz (@ecekt2)Photo by Ece Takmaz (@ecekt2)Photo by Ece Takmaz (@ecekt2)Photo by Ece Takmaz (@ecekt2)Photo by Ece Takmaz (@ecekt2)Photo by Ece Takmaz (@ecekt2)
Ece Takmaz retweeted
Thank you to everyone who helped make the first MINT workshop at #EMNLP2026 such a success! 🌿 We hope it’s the first of many MINTs 🌿🌿🌿! The proceedings are available in the ACL Anthology: aclanthology.org/volumes/202… #NLP #AI #multimodal_learning #language_modelling
1
1
3
170
Ece Takmaz retweeted
For those at EMNLP: tomorrow at 9:50 I will give a keynote talk at the MRL workshop about my work on multilingual linguistic evaluation. Artisanal, Claude-free slides guaranteed, for those in need of a palette cleanser after the past few days!
1
4
43
1,352
Ece Takmaz retweeted
New paper at babylm.github.io/ on "Relational Attention for Data-Efficient Language Modeling" with @ecekt2 and Jakub Dotlačil, building on the seminal work of Taylor Webb, @awni_altabaa and others. Paper: arxiv.org/abs/2609.20530. Code: github.com/abrsvn/babylm_dat….
1
1
1
9
3,345
Submit to BabyVLM @NeurIPSConf : babyvlm.github.io/ We welcome work on: 🧠 Sample-efficient learning 👀 Vision + language 🌍 Grounded learning 👶 Developmental learning 📏 Human-inspired evaluation 🤖 Learning from limited data 📅 Deadline: August 31
2
10
676
This week I’m in Utrecht! I’m visiting Utrecht University to give an invited seminar talk on pixel language models, hosted by the NLP Research Group led by Albert Gatt. Note: Don’t be fooled by the image; the weather is cloudy as expected! #NLProc
1
15
985
For people thinking that DeepSeek-OCR is the first model to render text as images, the University of Copenhagen already did this in 2023 Paper is called "Language Modelling with Pixels". They trained a Masked AutoEncoder (MAE) by rendering text as images and masking patches
22
49
7
512
45,007
Thanks Niels! It's cool to see bigger players getting into this topic. We've also done follow-up work on downstream applications and multilingual pretraining lampgroup.github.io/projects…. @mdlhx's group at KU Leuven has also done cool work on pixel evaluation lagom.cs.kuleuven.be/
For people thinking that DeepSeek-OCR is the first model to render text as images, the University of Copenhagen already did this in 2023 Paper is called "Language Modelling with Pixels". They trained a Masked AutoEncoder (MAE) by rendering text as images and masking patches
4
19
2,405
🌍Introducing BabyBabelLM: A Multilingual Benchmark of Developmentally Plausible Training Data! LLMs learn from vastly more data than humans ever experience. BabyLM challenges this paradigm by focusing on developmentally plausible data We extend this effort to 45 new languages!
1
19
4
42
4,577
Now I humbly request your participation May 3 in Santa Ana @naaclmeeting; @ttk_kuribayashi and I will be there
📣 We are happy to share that CMCL 2025 will be co-located with NAACL in New Mexico! 👉The call for papers is out cmclorg.github.io/CfP SAVE THE DATES, and submit your work! ‼️ Paper submission deadline: Feb 16, 2025 🗓 May 3 or 4, 2025: Workshop (TBA) @naacl @naaclmeeting
1
1
3
3,154
I'm attending #NAACL2025 to organize the CMCL workshop in person (May 3rd, Santa Ana room) w/ @byungdoh! We received many insightful cognitively motivated NLP works this year, too. Stay tuned! Program: cmclorg.github.io/program
Now I humbly request your participation May 3 in Santa Ana @naaclmeeting; @ttk_kuribayashi and I will be there
3
16
2,445
Ece Takmaz retweeted
Abstract submissions for the 1st Computational Psycholinguistics Meeting is now open! cpl2025.sites.uu.nl/
1
10
1
27
3,579
I’m really sad that my dear friend @FelixHill84 is no longer with us. He had many friends and colleagues all over the world - to try to ensure we reach them, his family have asked to share this webpage for the celebration of his life: pp.events/felix
109
81
79
716
302,127
Replying to @_philschmid @OpenAI
Not to confuse with JUDGE-BENCH, with its 85K instances and 20 NLP tasks ;) Paper: arxiv.org/pdf/2406.18403 Github: github.com/dmg-illc/JUDGE-BE…
3
14
755
Ece Takmaz retweeted
⚠️Urgently looking for emergency reviewers for the Multimodal NLP and Grounded Language Acquisition, HRI track in COLING 2025! Let me know if you would like to help (reviews to be submitted ideally before October 29). #nlproc
1
3
7
2,824
✨Want to find out about and discuss promising future directions in ML research? ✨Tune into our upcoming Research Connections social @CohereForAI's community, today at 1:30 pm EST! We're excited to host Ece Takmaz ( @ecekt2 ), Alon Albalak ( @albalakalon ), and Qiwei Peng 🎉
1
5
1
10
3,273
Very interesting paper by @HuaYilun and @yoavartzi! Great to see work on this topic dear to me: adaptation in repeated reference games. We looked into it using earlier models in this EMNLP-2020 paper aclanthology.org/2020.emnlp-… with @ecekt2 @glnmario @sandropezzelle @janiesinclair
Humans rapidly develop more efficient language and understand each other better as they interact. Do multimodal LLMs (MLLMs) pick up this ability for rapid ad-hoc adaptation from their training?
5
22
7,515
#CMCL2024 has come to an end, but we are already looking forward to #CMCL2025! Follow us on X, we will give you updates for news of the following months! @ttk_kuribayashi @ecekt2 @g_rambelli @jixingli
3
8
1,195
📣 The last speaker of the day is @aidanematzadeh with her work on Leveraging Cognitive Science to Unravel the Complexities of Generative Models 💡 even the long day most of the attendees are still there!💪 #CMCL2024 #CMCL #ACL2024
5
21
1,385
📣 Our second invited speaker is on the floor! Frank Keller (University of Edinburgh) is speaking on predicting sentence production across vision and language 🧐 #CMCL #CMCL2024 #ACL2024
3
7
800