Edoardo Ponti
@edoardo-ponti.bsky.social
Assistant professor in Natural Language Processing at the University of Edinburgh and visiting professor at NVIDIA | A Kleene star shines on the hour of our meeting.
Up next on stage, Dr. @edoardo-ponti.bsky.social ( @edinburgh-uni.bsky.social / NVIDIA) 🎤 “Adaptive Units of Computation: Towards Sublinear-Memory and Tokenizer-Free Foundation Models” Fascinating glimpse into the next gen of foundation models. #FoundationModels #NLP #TokenizerFree #ADSAI2025
🚀 By *learning* to compress the KV cache in Transformer LLMs, we can generate more tokens for the same compute budget. This unlocks *inference-time hyper-scaling* For the same runtime or memory load, we can boost LLM accuracy by pushing reasoning even further!
We propose Neurosymbolic Diffusion Models! We find diffusion is especially compelling for neurosymbolic approaches, combining powerful multimodal understanding with symbolic reasoning 🚀 Read more 👇
Sparse attention is one of the most promising strategies to unlock long-context processing and long-generation reasoning in LLMs. We performed the most comprehensive study on training-free sparse attention to date. Here is what we found:
🚀 Excited to welcome Dr. @edoardo-ponti.bsky.social to #ADSAI2025! Lecturer in NLP @edinburghuni.bsky.social , Affiliated Lecturer @cambridgeuni.bsky.social & Visiting Prof NVIDIA. 🎟️ Tickets for Advances in Data Science & AI Conference 2025 are live! 🔗Secure your spot: tinyurl.com/yurknk7y #AI
We created Approximate Likelihood Matching, a principled (and very effective) method for *cross-tokenizer distillation*! With ALM, you can create ensembles of models from different families, convert existing subword-level models to byte-level and a bunch more🧵
I have a scholarship for a PhD in efficient memory and tokenization in LLM architectures at @edinburgh-uni.bsky.social! Eligibility: UK home fee status Starting date: flexible, from July 2025 onwards. informatics.ed.ac.uk/study-with-u... Please contact me if you're interested!
Code and models for Dynamic Memory Compression are finally available! Stay tuned for architectures with even more efficient inference. developer.nvidia.com/blog/dynamic...
Dynamic Memory Compression | NVIDIA Technical Blog
Despite the success of large language models (LLMs) as general-purpose AI tools, their high demand for computational resources make their deployment challenging in many real-world scenarios.
developer.nvidia.com
We're hiring a lecturer or reader in embodied NLP at the University of Edinburgh! Deadline: 31 Jan 2025 Call for applications: elxw.fa.em3.oraclecloud.com/hcmUI/Candid...
**Grounded typology**: a new paradigm. Traditionally, linguists posit functions to compare forms in different languages; however, these are aprioristic and partly arbitrary. Instead, we resort to perceptual modalities (like vision) as measurable proxies for function.
Coleman Haley@colemanhaley.bsky.social · 2y agoNEW PREPRINT! Language is not just a formal system—it connects words to the world. But how do we measure this connection in a cross-linguistic, quantitative way? 🧵 Using multimodal models, we introduce a new approach: groundedness ⬇️
Two amazing papers from my students at #NeurIPS today: ⛓️💥 Switch the vocabulary and embeddings of your LLM tokenizer zero-shot on the fly (@bminixhofer.bsky.social) neurips.cc/virtual/2024... 🌊 Align your LLM gradient-free with spectral editing of activations (Yifu Qiu) neurips.cc/virtual/2024...
We had a blast at this year's @ellis.eu Dagstuhl seminar on "Modular and Agentive LLMs". Thanks everyone for participating!
Check out this piece on Strawberry 🍓/o1 we just authored on TheConversation! theconversation.com/ai-that-mimi... with @edoardo-ponti.bsky.social and Kolya 🚀
AI that mimics human problem solving is a big advance – but comes with new risks and problems
Chain of thought reasoning has been used in OpenAI’s new AI model.
theconversation.com
Several incredible NeurIPS tutorials this year. Worth navigating through the Swifties.
Edoardo Ponti@edoardo-ponti.bsky.social · 2y agoAnother nano gem from my amazing student Piotr Nawrot! A repo & notebook on sparse attention for efficient LLM inference: github.com/PiotrNawrot/... This will also feature in my #NeurIPS 2024 tutorial "Dynamic Sparsity in ML" with André Martins: dynamic-sparsity.github.io Stay tuned!
P.S. Make sure to follow @pnawrot.bsky.social!
Edoardo Ponti@edoardo-ponti.bsky.social · 2y agoAnother nano gem from my amazing student Piotr Nawrot! A repo & notebook on sparse attention for efficient LLM inference: github.com/PiotrNawrot/... This will also feature in my #NeurIPS 2024 tutorial "Dynamic Sparsity in ML" with André Martins: dynamic-sparsity.github.io Stay tuned!
Explore zero-shot routing of parameter-efficient experts with Phatgoose arxiv.org/abs/2402.05859 and Arrow arxiv.org/abs/2405.11157 w. github.com/microsoft/mttl 👉 github.com/sordonia/pg_mb… Part of "Dynamic Sparsity in ML" tuto #neurips2024, feedback welcome and join for discussions! 😊
Last 5 days to apply for a PhD at #EdinburghNLP! Deadline: November 25 www.ed.ac.uk/studying/pos... If you are passionate about: - adaptive tokenization and memory in foundation models - modular deep learning - computational typology please message me or meet me at #NeurIPS2024!
Informatics: ILCC: Language Processing, Speech Technology, Information Retrieval, Cognition
Study Informatics: ILCC: Language Processing, Speech Technology, Information Retrieval, Cognition at the University of Edinburgh. Our postgraduate degree programmes focus on natural language processin...
ed.ac.uk
Another nano gem from my amazing student Piotr Nawrot! A repo & notebook on sparse attention for efficient LLM inference: github.com/PiotrNawrot/... This will also feature in my #NeurIPS 2024 tutorial "Dynamic Sparsity in ML" with André Martins: dynamic-sparsity.github.io Stay tuned!