Reduce LLM hallucinations with RAG over textual as well as structured knowledge bases. Together with Amazon we are releasing 🌟STaRK 🌟, a large-scale LLM retrieval benchmark on semi-structured knowledge bases with dataset from e-commerce, biomedicine, and academic research.

Research on Slovenian 🇸🇮 cultural heritage in the USA 🇺🇸 and Canada 🇨🇦 is extended until 𝗠𝗮𝘆 𝟭𝟱, 𝟮𝟬𝟮𝟰❗

🔗 Survey for churches/parishes ➡️ bit.ly/surveyparishes
🔗 Survey for other organizations/structures ➡️ bit.ly/surveyorganiza…

Kumo.AI has just been recognized by Forbes as one of the top AI companies in the world alongside giants like OpenAI and Anthropic.
I couldn't be prouder to be part of the Kumo.AI and the AI revolution.
forbes.com/lists/ai50/?sh…

What a way to start the day! Kumo.AI has just been recognized by Forbes as one of the top AI companies in the world. It feels surreal to be ranked alongside giants like OpenAI and Anthropic.

I couldn't be prouder to be part of the Kumo.AI and the AI…

Check out our PyTorch Frame tech report arxiv.org/abs/2404.00776

We aim to push tabular deep learning to handle complex multi-modal, multi-tabular data, beyond conventional single-tabular data. We do so by seamlessly integrating PyTorch Frame with LLMs and Graph Neural Nets!

From Yanay Rosen and Jure Leskovec : A foundation model that universally maps cells across tissues and species! Learn more at : github.com/snap-stanford/…

Thrilled to share that my lab is looking for postdocs! In partnership with Chan Zuckerberg Initiative, we're focusing on developing massive biomedical foundation models to create an AI-powered virtual cell. Dream of harnessing the power of 1,000 H100s? Apply now at: snap.stanford.edu/apply/index-po…

Grateful to have sat down with Carl for this story to share more about finding from UCE, which was drained on data from CZI Science's #CellxGene . Thank you to collaborators Yanay Rosen Yusuf Roohani and CZI!

Excited to share our PlaNet 🌏method that reasons over population variability, disease biology, and drug chemistry using a massive clinical knowledge graph. We apply PlaNet to predict outcomes of clinical trials.

w/ Michi Yasunaga Prabhat Agarwal Jure Leskovec
medrxiv.org/content/10.110…

🚀 Join us for a special webinar on March 6th, 8am PT/5pm CET, as we unveil the latest in GNN technology - PyG 2.5! 🎉 Dive deep into the features with a live demo on setting up and using our new distributed GNN training solution. A stellar collaboration between Kumo.AI…

Our work on SATURN, a method that brings us closer to Universal Cell Embeddings across species has been published in Nature Methods today.

📢 In collaboration with Intel and Kumo.AI we are excited to release PyG v2.5.
Big new features: Distributed GNN training, graph tensor representation, RecSys support, PyTorch 2.2 and native compilation support!
github.com/pyg-team/pytor…

🚨1/ New to CZ #CELLxGENE : models & embeddings that integrate up to 36M cells in the Census corpus.

Use embeddings to explore the corpus directly, or download the models to run your own data through them to enable direct comparisons to the reference. 🧵

bit.ly/census_models

Nice work from Kexin Huang and collaborators at Genentech that trained GEARS in the active learning setting using iterative wet-lab experiments

Making tabular machine learning easy with PyToch Frame. Check it out!

Can deep learning work on small data with far more features than samples? We present PLATO: a method that achieves the state-of-the-art on such datasets by using prior domain information! neurips.cc/virtual/2022/p… 🧵

Published in #NeurIPS2023 with Hongyu Ren @kexinhuang5 Jure Leskovec

Unrelated to all the goings on at NeurIPS, this is now available on arXiv!

Paper: arxiv.org/abs/2312.04615

Excited to share our latest research: The myth of cosmopolitan cities: Why large urban areas are more segregated

Long-standing assumption is that large cities with their diverse population, foster diverse person-to-person interactions.

Our nature paper shows the opposite is…

Relational Deep Learning is bringing the power of representation learning to relational databases and data warehouses. This breakthrough technology offers lots of opportunity for impact and raises exciting new research questions.
Check out the paper here: relbench.stanford.edu/paper.pdf…