Our Mission


To usher in a new class of machine learning for scientific data, building models that can leverage shared concepts across disciplines. We aim to develop, train, and release such foundation models for use by researchers worldwide.

Read More

Recent News


Article Blog Image

Counting in Context?

May 30, 2024

We introduce the Contextual Counting task, a new toy problem aimed at exploring interpretability of Transformer models in quantitative domains.

Article Blog Image

xVal Number Encoding

Oct 09, 2023

We introduce xVal, a new number encoding scheme for LLMs. Using xVal with a modified number inference method makes LLMs continuous function approximators. This makes them have a better inductive bias for data analysis in scientific domains.

Article Blog Image

Learning Multiple Physics

Oct 09, 2023

Introducing Multiple Physics Pretraining, a new approach for developing large pretrained physical surrogate models. Our approach uses a built-in normalization and embedding scheme to enable learning multiple physical dynamics with a single model.

The Team


Scientific Advisory Group


Colm-Cille
Caulfield

University of Cambridge

Leslie
Greengard

Flatiron Institute
New York University

David
Ha

Sakana AI

Yann
LeCun

Meta AI
New York University

Stephane
Mallat

École Normale Supérieure
Collège de France
Flatiron Institute

David
Spergel


Simons Foundation
 

Olga
Troyanskaya

Flatiron Institute
Princeton University

Laure
Zanna

New York University

Participating Institutions