Transformer jointly models electrons and magnetism with 0.01% energy error
bravo_abad · x · 2026-10-23
Rende and colleagues build electron–spin interactions directly into a transformer-based neural quantum state (NQS) to simulate electrons in magnetic materials.
Key design choices:
- Each lattice site becomes one token carrying both electron occupation and local spins, avoiding a rigid separation of the two physics
- The transformer generates electronic orbitals — the building blocks of the wave function — from the full configuration, and orbitals can change as the spin configuration changes
- The architecture figure shows the full pipeline: tokenization → embedding → multi-layer transformer → orbital matrix → wave function via determinant, Ψθ(s) = det[n ★ Φ(s)]
The joint modeling of electron–magnetism coupling reaches 0.01% energy error, showing NQS promise for strongly correlated quantum systems.
More from Research
- DeepMind's IPO Paper Gets a 1000x Citation 'Hack' — Thanks to Another Paper Named IPO — _lewtun · 2026-10-23
- Why LLMs dissolve the long-standing barriers to formally verified software — teortaxesTex · 2026-10-23
- Japan's Moonshot IoB Brain Project Publishes 5-Year Research Posters Online — kaixhin · 2026-10-23
- Agent evals should test metric definition drift, argues dev in new post — jylusdev · 2026-10-23
- ruff author asks: how hard is scaling AI chip design from RISC-V to x86? — charliermarsh · 2026-10-23
- Google Cloud AI releases PLE benchmark; all frontier models score under 50% — sainingxie · 2026-10-23