🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Check out our latest research work "Rad-DINO: Exploring Scalable Medical Image Encoders Beyond Text Supervision" arxiv.org/pdf/2401.10815… We investigate whether text supervision (e.g. CLIP) is an essential component for learning rich visual embeddings to encode biomedical images

Rᴀᴅ-DINO is now on arXiv!🦖 We show that medical image encoders trained only on images perform similar or better than text supervised models on various benchmarks, demonstrating that reliance on text may not be necessary, and can become a potential limitation. Microsoft Research

🔥 Our new preprint on biomedical self-supervised learning with images only is out! 🔥 Rᴀᴅ-DINO: Exploring Scalable Medical Image Encoders Beyond Text Supervision by Microsoft Research We show that: 0/3 🧵

Considerable work aims to mitigate 'algorithmic bias', assumed to be inherited from training data. However, little work stops to consider the underlying structures of dataset bias... Our latest Nature Machine Intelligence Perspective provides a unifying causal view on dataset bias. (1/5)

Amit Sharma Thanks for the great thread! You may also be interested in our #ICLR2024 paper, where we demonstrate that combining (1) LLM causal hypothesis generation & critique with (2) grounding on evidence from real data greatly improves causal discovery. openreview.net/forum?id=pAoqR…

Vision-Language Models will revolutionize radiology and enhance patient care. How do we ensure they achieve their goals in practice? Latest research from Microsoft Research led by @runmiridliy explores clinician+AI interaction for VLMs in healthcare! lnkd.in/gQv6RNq4

Thrilled to announce our collaboration with Cyted Health and Cancer Research UK to scale early detection of esophageal cancer with AI.

🔬 Optimising pathology workflows for early detection of oesophageal cancer, using large-scale ML models for whole-slide image analysis. Many interesting clinical, modelling, and engineering challenges, working in close collaboration with stakeholders. Proud to see it published!

It's great to see our work on causality in fair machine learning showcased in the Nature Portfolio collection on AI and robotics! nature.com/immersive/robo…

We are hiring a senior researcher in ML for healthcare at MSR Cambridge (UK)! The position is in my team, so if you get it you will work with me (is this a pro or a con? do not answer). Focus is multimodal (~vision-language) models for radiology! Link: jobs.careers.microsoft.com/global/en/job/…

We've released the Rᴀᴅ-DINO model weights!! Benchmark it, encode some datasets, show us some UMAPs, plug it into your classifiers, LLMs, MLMs, SLMs... We're excited to discover what the community will create on top of Rᴀᴅ-DINO. 🤗aka.ms/rad-dino-model Microsoft Research

my team's radiology report generation metric is now open source: microsoft/RadFact: A metric suite leveraging the logical inference capabilities of LLMs, for radiology report generation both with and without grounding (github.com)#MoreMetrics arxiv.org/abs/2406.04449

RadEdit stress-tests biomedical vision models by simulating dataset shifts through precise image editing. It uses diffusion models to create realistic, synthetic datasets, helping to identify model weaknesses and evaluate robustness: msft.it/6011mlGU3

THIS X-RAY DOESN'T EXIST European Conference on Computer Vision #ECCV2026 #ECCV2024 🧵1/9

Our MAIRA-2 model for (grounded) radiology report generation from CXRs is out now! huggingface.co/microsoft/mair…

if you find yourself wanting to understand how and why models work, in a way that could be useful for biomedical discovery, come to Cambridge (UK) for a postdoc at MSR: jobs.careers.microsoft.com/global/en/job/…

RAD-DINO: Exploring Scalable Medical Image Encoders Beyond Text Supervision, published today in Nature Machine Intelligence. 👉 Link nature.com/articles/s4225… #microsoft #healthcare #ai #ResearchPapers

Today, I am excited to share our collaboration with Mayo Clinic. newsnetwork.mayoclinic.org/discussion/may… Our collaboration brings together the best experts from AI and medicine to unlock new frontiers in radiology AI #JPM25

The foundation models we'll build with Mayo Clinic promise to help transform how radiologists do their work, using multimodal AI to help them analyze X-rays faster and more accurately.

From generating novel materials with MatterGen to helping to improve diagnostic accuracy with RAD-DINO, Microsoft’s latest AI models are set to transform the fields of science & healthcare. msft.it/6003UDFZF