Very pleased to share our latest paper, and the first by our student Tonny Lou, showing how the FermiNet can be used to study the Fermi gas, a model system for superconductors and superfluids! arxiv.org/abs/2305.06989 The Fermi gas is a model system in condensed matter physics that falls about halfway between a Bose-Einstein condensate and a conventional superconductor. It’s made out of electrons, but electrons of opposite spin have a local *attractive* force, unlike the usual Coulomb force.

And there you have it. 2 MJ laser energy in, 3 MJ fusion energy out. A sputnik moment for fusion research. The shot heard round the world.

energy.gov/articles/doe-n… These milestones can often be hard to parse, because the definition of "energy in" can vary quite wildly for inertial (laser) confinement fusion.

OK, this website still seems to be working…so…time to share our latest preprint!

Very pleased to be able to share this one: is attention all you need to solve the Schrödinger equation? arxiv.org/abs/2211.13672 For the last several years, numerous groups have shown that neural networks can make calculations in quantum chemistry much more accurate - FermiNet, PauliNet, etc. We wrote a review article about it here:

https://twitter.com/jhrmnn/status/1564152388491132928

@KordingLab Continue? It never drove it in the first place. @KordingLab Do you think the people who came up with Transformers ever read a single neuroscience paper? Or ADAM? You're all just talking to yourselves and pretending you're contributing. It's embarrassing.

OK, this is one I’ve been waiting to share for a *long* time – the first ever demonstration of deep reinforcement learning on a nuclear fusion research device! nature.com/articles/s4158… Nuclear fusion is the holy grail of clean energy - abundant fuel, small footprint, runs 24/7, zero meltdown risk or long-lasting waste. But despite 70 years of work, it has yet to become a reality. It is more familiar from science fiction than real life.

2014: Integrated information theory can't be correct - if it were, matrix multiplication would be conscious.

2021: Maybe matrix multiplication is conscious? For context (at least for the first one): scottaaronson.blog/?p=1799

Alright, if we're doing U.F.O. discourse now, I'm going to place my bet on what's going on: it's some kind of coherent plasma structure shot out of the Sun as part of a coronal mass ejection.

nytimes.com/2021/06/03/us/… One of the most well-documented U.F.O. sightings in recent years was the "Tic Tac Incident" of November 10, 2004. You can read about it here: nymag.com/intelligencer/…

Bold move by Greene and Gosar saying we should overturn hundreds of years of development of the American legal system and replace it with the full English common law system. talkingpointsmemo.com/news/new-gop-c… Getting Louisiana to finally give up the Napoleonic Code is definitely the winning issue for the GOP to rally around

Excited to share our latest work: the Geometric Manifold Component Estimator, or GEOMANCER, a nonparametric algorithm for symmetry-based disentangling! 1/n

Paper: arxiv.org/abs/2006.12982
Code: github.com/deepmind/deepm… "Disentangling" is a somewhat nebulous term in ML, but it is broadly about building models that can separate out different latent factors of variation - for instance, in vision, separating translation, rotation, and changes in lighting or color that leave objects invariant. 2/n

Thrilled to be able to share what I've been working on for the last year - solving the fundamental equations of quantum mechanics with deep learning!

arxiv.org/abs/1909.02487 The Schroedinger equation - basically Newton's laws at the atomic scale - have been known for almost 100 years. But the equations are impossible to solve in closed form for anything more complicated than a hydrogen atom.