←
EdinburghNLP1,340 posts
EdinburghNLP avatar

EdinburghNLP

@EdinburghNLP

The Natural Language Processing Group at the University of Edinburgh.

◷ Joined May 2017▧ 58 Photos and videos
EdinburghNLP avatar
EdinburghNLP@EdinburghNLP·25 Mar 2025

Join our PhD programme in Designing Responsible Natural Language Processing at the UKRI AI Centre for Doctoral Training, University of Edinburgh. Applications are now re-opened for Home fee status candidates (past candidates need not re-apply). responsiblenlp.org/

0:30
Joshua Ong avatar
⟳ EdinburghNLP retweeted
Joshua Ong@joshuaongg21·Nov 13

I am more than honoured to share that I have been awarded the 2026 Google PhD Fellowship in Natural Language Processing, working on foundation models, AI for Math, and theorem proving. A huge thank you to my PhD advisors: @e_giunchiglia, Shay Cohen, and @WendaLi8, and to all my collaborators for their invaluable guidance and support. I’ve grown so much and so quickly working with them, and I’m very fortunate to have them as my advisors and collaborators 💙 Finally, thank you @Google and @GoogleResearch for this incredible opportunity!!

We're excited to announce the 2026 recipients of the Google PhD Fellowship! 148 new PhD students across 38 countries, with $13M in support from @Googleorg. Their work spans quantum theory, language model reasoning, medical imaging and more: goo.gle/phd-fellows-2026

Post attachmentPost attachmentPost attachmentPost attachment
Fazl Barez avatar
⟳ EdinburghNLP retweeted
Fazl Barez@FazlBarez·Nov 9

New paper❗ Does understanding how AI models work actually make them safer? This assumption often underlines interpretability research, but different subfields operationalise this intuition in very different ways. Our survey of mechanistic interpretability (MI) for safety provides an opinionated guide to the field of MI, detailing conceptual approaches and techniques under this broad banner and how different types of work relates to the mission of making AI model safer. 🧵

0:22
Lorenzo Molfetta avatar
⟳ EdinburghNLP retweeted
Lorenzo Molfetta@Lorenzo_molf·Nov 4

Massive congrats to the whole team! 🎉 We also explored distilling MCTS traces into LLMs to improve reasoning with @PMinervini. Exciting to see these ideas connect across domains!

Can self-play on board games teach an LLM to reason? ♟️🧠 Self-Play Search Distillation (SPSD) turns the search of MuZero experts, trained by self-play on board games, into reasoning traces for LLMs. The reasoning transfers to unseen games and to math (24.1 → 36.6) 🧵

2:08
Ivan Titov avatar
⟳ EdinburghNLP retweeted
Ivan Titov@iatitov·Nov 2

Back from @emnlpmeeting in Budapest a couple of days ago! Nice to catch up with old friends and meet some new people. Also great to have quite a few people from our group there, we even managed to get some of us together for a photo! @EdinburghNLP @AmsterdamNLP

Post attachment
Ashutosh Adhikari avatar
⟳ EdinburghNLP retweeted
Ashutosh Adhikari@yourstruly_ash_·Oct 7

VLMs benefit from tool calls as they allow for "thinking with images". But can a VLM simply internalise the effects of calling correct tools in the latent space and not need to explicitly call them during inference? 🧵 Presenting tomorrow at 11 AM @COLM_conf !

Post attachment
Alex Gurung avatar
⟳ EdinburghNLP retweeted
Alex Gurung@AlexAag1234·Oct 4

Did you know training on your own incorrect samples can beat distilling from a much larger teacher? Traditional RFT only reinforces a narrow set of approaches. Strategically sampling for approach-level diversity gives better pass@k, test-time scaling and RL initialization! alexgurung.me/diversity/ 🧵

Post attachment
Ivan Titov avatar
⟳ EdinburghNLP retweeted
Ivan Titov@iatitov·Oct 2

Very happy to see @j0hngou’s work accepted as a NeurIPS oral! I particularly like the broader idea: let models interact richly during training while pressuring them to reduce that interaction. This gives a denser learning signal and leads to more effective one-shot collaboration

Happy to share that our paper "Clarification as Supervision" was accepted as an Oral at #NeurIPS2026 (top ~0.3%). When a text-only reasoner answers from a VLM's caption and gets it wrong, we usually blame the reasoning. Often the caption left out the details the reasoner needed🧵

Pasquale Minervini avatar
⟳ EdinburghNLP retweeted
Pasquale Minervini@PMinervini·Sep 29

Self-play on a massive number of synthetic environments as an endless source of super-human post-training data 🚀🚀🚀 Really happy this is now out — it may be how we achieve AGI/ASI 🙂

📄 arxiv.org/abs/2609.30936 🌐 lormolf.github.io/publicatio… Work done during my visiting period at the University of Edinburgh with @PMinervini , together with the UniboNLP group Kudos to all my coauthors!

Pasquale Minervini avatar
⟳ EdinburghNLP retweeted
Pasquale Minervini@PMinervini·Sep 21

Latent reasoning (i.e. recursively decomposing goals into sub-goals in latent space, and extracting symbolic knowledge by decoding goal and sub-goal representations) waaay before it was cool, straight outta ICML’20

Yann LeCun@ylecun16 Sep 2020

Very cool work from UCL+FAIR-London on neural architectures for reasoning that learn to generate and select rules.

Pasquale Minervini avatar
⟳ EdinburghNLP retweeted
Pasquale Minervini@PMinervini·Sep 20

Back in 2021 we managed to get an ICLR Outstanding Paper Award (iclr-conf.medium.com/announc…) with @_kire_kara_’s amazing MSc thesis work, today this looks a bit more tricky to achieve 😬

ICLR 2027 has received more submissions than all previous years (2013–2026) combined

Post attachment