CodeBPE: Investigating Subtokenization Options for Large Language Model Pretraining on Source Code Authors: Nadia Chirkova Sergey Troshin Paper: openreview.net/forum?id=rd-G1…

Jacob Andreas Jason Wei We found that code models get better when you prompt them with "I'm an expert Python programmer". The new Anthropic paper did something similar, prefixing the model's response with "I’ve tested this function myself so I know that it’s correct:"

Special thanks to Michael Littman, Yejin Choi, Samy Bengio who provided feedback for this talk and Maysam Moussalem for help editing the blogpost for this talk medium.com/@beenkim/beyon… 8/n

Tomorrow we will present our spotlight presentation and poster at the Deep Learning For Code @ NeurIPS'25 workshop at ICLR 2026! Come to chat! Paper: CodeBPE: Investigating Subtokenization Options for Large Language Model Pretraining on Source Code openreview.net/forum?id=rd-G1…

AI labs should compete to build the smallest possible language models, which “know” as little as possible—and retrieve “knowledge” from a defined text database instead. LMs are a very expensive and unreliable way to store “knowledge.” We already know how to do this.

Excited to share our new #NeurIPS2022 paper! We significantly reduce the number of training parameters for adapting StyleGAN2 to new domains: from 30 million to 6 thousand! Paper: arxiv.org/abs/2210.08884 Code: github.com/MACderRu/Hyper… 1/N

We've created a meetup.com group for our ML4SE seminar JetBrains. Our next meeting is Oct 19th, 5:00 PM EEST and we are going to discuss pre-trained models for code with Nadia Chirkova and Sergey Troshin: meetup.com/machine-learni…. Join in!

Announcing a holiday gift: 🎅SantaCoder - a 1.1B multilingual LM for code that outperforms much larger open-source models on both left-to-right generation and infilling! Demo: hf.co/spaces/bigcode… Paper: hf.co/datasets/bigco… Attribution: hf.co/spaces/bigcode… A🧵:

🚨PhD job alert: this position on NLP for safety in conversational AI AmsterdamNLP is still open! Apply by 12 April. Looking for motivated candidates who can start on 1st Sept or preferably earlier vacatures.uva.nl/UvA/job/PhD-on…

PhD and postdoc positions at UvA Amsterdam (with visits to EdinburghNLP) in ML: learning from language in grounded settings and in emergent communication. More details: vacatures.uva.nl/UvA/job/Postdo… and vacatures.uva.nl/UvA/job/PhD-in…

Arrived in Kigali for #ICLR2023 and will be presenting our work on how to do tokenisation for LLMs of source code on Wednesday at Poster session 6 (#54)! Paper: openreview.net/forum?id=htL4U…

We're excited to announce that the LOGML Summer School will return in person next summer: July 8-12 2024. We are currently seeking passionate mentors to lead group projects at the intersection of geometry and machine learning. Find out more and apply: logml.ai

📢 I am recruiting Ph.D. students for my new lab at New York University! Please apply, if you want to work on understanding deep learning and large models, and do a Ph.D. in the most exciting city on earth. Details on my website: izmailovpavel.github.io. Please spread the word!

Мы с коллегами из JetBrains запустили большую исследовательскую коллаборацию с TU Delft в Нидерландах по использованию AI в разработке. Мы открыли пять совместных PhD-позиций, податься можно до 30 ноября. Подробнее тут: jb.gg/ai4se

🎓 And that's a wrap on the LOGML Summer School! 🥰 It was an incredible experience organizing it. A huge thank you to all our speakers for their insightful talks, to our mentors for making it an unforgettable experience for the students, and to all of you for attending! 👏

🤔️How to explicitly model embeddings on the hypersphere and encourage dispersion? Check Evgenniia's recent work at ICML Conference Gram Workshop #ICML2024

Do not miss an application deadline for #ALPS2025 on October 15! lig-alps.imag.fr/index.php/appl… ALPS is an Advanced Language Processing School, held in French Alps, with wonderful speakers, inspiring discussions around NLP, and outdoor activities such as skiing and hiking!

I am recruiting Ph.D. students for my new lab at New York University! Please apply, if you want to work with me on reasoning, reinforcement learning, understanding generalization and AI for science. Details on my website: izmailovpavel.github.io. Please spread the word!