Very excited about this work by our joint team. There is much more to explore in domain-specific pretraining. We also hope this work would help lower the entry barrier to biomedical NLP and accelerate progress in this vitally important area.

We are excited to announce the first group of #OxML 2022 speakers, including some of the world's top scientists in the field of #MachineLearning, from top #AI institutions/universities 👏 Application deadline is 15 April (oxfordml.school) CIFAR, Deep Medicine,

Hello #causaltwitter! We're hosting a 2-hr community workshop on May 3 bringing together folks working w/causal methods across a variety of industries and domains, with fireside chats; updates on our DoWhy, EconML and other tooling; and deep dives. microsoft.com/en-us/research…

Hi! I am looking for two postdocs (2 years) to work on causal inference and related topics in the healthcare domain. We'll be working with Hoifung Poon in some really cool and interesting projects. Send me a PM or apply here if you are interested. careers.microsoft.com/us/en/job/1171…

Gaussian process summer school 2022 will be in person, 12-15 Sept in Sheffield. Registration now open: gpss.cc/gpss22/ Speaker list still to be finalised, but expect the usual great list of engaging speakers. Neil Lawrence Mike Smith @maalvarezl Javier González wil o. c. ward

Interested in biomedical LLMs or multi-modal learning for precision health? We Microsoft Research Health Futures are too! lnkd.in/gsjv-eJd

Heading to KDD... Excited to present closing keynote tomorrow at dshealthkdd.github.io/dshealth-2023/ and join colleagues Sheng Zhang Javier González Tristan Naumann on Thu 8/10 to present tutorial "Precision Health at the Age of LLMs". Also happy to discuss career opportunities (we're hiring!).

Getting ready for KDD! With Sheng Zhang, Tristan Naumann and Hoifung Poon we’ll be giving a tutorial on “Precision Health in the Age of LLMs” on August 10th. We are also hiring in Cambridge, UK. Check here for available opportunities: jobs.careers.microsoft.com/global/en/job/…

Our paper "Beyond Words: A Mathematical Framework for Interpreting Large Language Models" is now available here: arxiv.org/abs/2311.03033. We introduce a novel framework, Hex, that helps clarify and formalize key terms and concepts in LLM research. Hope you like it!

I am having a fantastic time at the FIORD2023 workshop in Washington where I have presented or work 'TRIALSCOPE: A Unifying Causal Framework for Scaling Real-World Evidence Generation with Biomedical Language Models. arxiv.org/abs/2311.01301.

Let me call your attention to a new paper by my colleagues Javier González and Aditya Nori. Hex is a mathematical framework to formalize terms in LLMs. These are some concepts that can be defined mathematically with Hex: 🔹Alignment. 🔹Prompting. 🔹Hallucination.

We are hiring at MSR Cambridge to work in ML/GenAI on a variety of really exciting topics. Reach out if you are interested!

Economists seem to LOVE synthetic control methods, so during my MSR internship with Javier González we wanted to understand whether we could use them for survival analyses (v prevalent in medicine) too? Delighted that our answer (“It’s complicated!”) was accepted CLeaR-Conference on Causal Learning and Reasoning… 1/n

Job opportunity Sheffield NLP 🚨: I'm looking for a #postdoc (24 months) in #NLProc The (very) broad topic is on addressing LLMs limitations (e.g. hallucinations, "reasoning", interpretability etc.) If you are interested drop me an email or DM Apply: jobs.ac.uk/job/DHP918/res…

Just out nature The 1st whole-slide digital pathology #AI foundation model pre-trained on large-scale real-world data, from over 1.3 billion images, 30,000 patients nature.com/articles/s4158… Hoifung Poon Hanwen Xu Microsoft University of Washington Naoto Usuyama Microsoft Research

We've made a ton of progress on medical imaging with AI. What is most important here is the possibility that there is strong possibility of useful scaling at play.

Whole-slide foundation model for pathology (Gigapath) paper is out in nature today - incredible work from Hoifung Poon Naoto Usuyama and the rest of the team

In nature today, our work on the first foundation model for whole-slide digital pathology. Key advance is it shows strong evidence of scaling benefit. State-of-the-art performance on 17/18 imaging tasks, with 12 showing very significant improvement. microsoft.com/en-us/research…

Excited to share that our GigaPath paper is published by Nature today: Paper: aka.ms/gigapath Model: aka.ms/prov-gigapath MSR Blog: microsoft.com/en-us/research…

We are thrilled to announce the release of our latest paper, “Does Reasoning Emerge? Examining the Probabilities of Causation in Large Language Models”, authored by Aditya Nori and myself. arxiv.org/abs/2408.08210 Hope you enjoy it!