Dr. Ruchira Datta

@ruchiradatta.bsky.social

Mathematician and computer scientist. Have done research in game theory, AI/ML for NLP, and computational biology (phylogenomics, cancer, and immunology). Now working on making energy-efficient AI/ML. Vegetarian.

"The central methodological implication is that, in climate finance, stability should be treated as an outcome to be investigated rather than an assumption imposed at the outset" These are "ruled out or strongly constrained in conventional equilibrium IAMs" academic.oup.com/imaman/advan...

Here we go again: the misuse of models and the looming climate crisis

Abstract. Climate-related financial risks are characterized by path dependence, non-linear feedbacks, regime shifts and the possibility of endogenous insta

academic.oup.com

Decode every shot twice: fast in real time, then with a slower, more accurate decoder afterward. Where they disagree, a logical error is likely. Discarding those shots still leaves a bias. Our new preprint keeps them, with negative weights, which cancels the bias on average.

Two panels. Compare two decoders: an offline decoder and a real-time decoder read the same syndromes; most shots agree, a small fraction disagree in X, Y or Z sectors. Reweight, don't discard: agreeing shots get a weight above one, disagreeing shots negative weights, moving the estimate from the biased raw value to the noiseless value.

DINOv2 has never seen a caption, and Qwen3 has never seen an image. We still aligned their embedding spaces without a single image-caption pair. It even works when the images and the captions come from different datasets. Project page: dominik-schnaus.github.io/unpaired-rosetta ⬇️

This is a result I’ve dreamt about for many years: *unpaired translation between images and text* I thought it might be only slightly possible, the kind of thing you have to really squint at. But Dominik proved this wrong. You don’t have to squint. Worth looking for yourself:

Dominik Schnaus@schnaus.bsky.social · 3d ago

DINOv2 has never seen a caption, and Qwen3 has never seen an image. We still aligned their embedding spaces without a single image-caption pair. It even works when the images and the captions come from different datasets. Project page: dominik-schnaus.github.io/unpaired-rosetta ⬇️

Rufous-Necked Hornbill one of the largest hornbill species, native to the Himalayas, northeast India, and parts of Southeast Asia Classified as Vulnerable on the IUCN Red List, with an estimated wild population under 10,000 mature individuals 📷: Kedar Potnis #birds

Bild