Papers for

environmental sustainability teams in tech

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Astronomy foundation model built from combined existing vision models

Gestalt: a meta-foundation model for astronomy

Abstract: The Platonic Representation Hypothesis predicts that sufficiently scaled foundation models converge on a shared representation of the world. As each non-converged model gives a noisy view of a common structure when passed the same input, we ask whether we can combine models into a representation that outperforms its individual components. We test this on galaxies: we embed images via a basket of 22 frozen foundation models from eight families, whiten each view, and take a randomised SVD of the embedding concatenation. The resulting 1024-dimensional embedding outperforms every basket member on 19/21 of our tested metrics for physical property and galaxy morphology estimation for HSC, JWST, and DESI Legacy Survey imagery. We find that performance rises with basket size and basket architectural diversity, and that the meta-foundation model's performance transfers across astronomical surveys. We conclude that a useful astronomical foundation model can be assembled from existing generalist models with no training required beyond a single unsupervised projection. By leveraging the community's already-spent work, we save a lot of compute: a fresh pre-train of a comparable single-domain model would cost $\mathcal{O}(10^{4}$--$10^{5})$ A100 GPU hours (emitting several tonnes of CO$_2$eq.), whereas assembling Gestalt requires minutes on a single machine.

Tue 29 SeptMachine Learning
The gist
Different big AI models can see the same galaxy image in different ways, like looking through slightly different glasses. The authors show that by combining 22 such models from different families, they can create a new model called Gestalt that understands galaxies better than any one model alone. This new model is fast to create and works well across different astronomical surveys without extra training. It saves a lot of computing power compared to training a new AI model from scratch.
Open → 2609.38312v1