Noé group releases transferable DeepQMC for excited-state potential energy surfaces in one run
FrankNoeBerlin · x · 2026-11-07
Frank Noé's group at Freie Universität Berlin (with Zeno Schätzle, P. Bernt Szabó, Alice Cuzzocrea, Matej Mezera) published a new paper on ab initio simulation of excited-state potential energy surfaces with transferable deep quantum Monte Carlo.
- Problem: Accurate quantum chemical excited-state calculations are hard; entire excited-state energy surfaces are super hard.
- Method: A transferable excited-state neural network wavefunction ansatz leveraging weight sharing and dynamical ordering of electronic states, optimizing multiple states in a single deep QMC run. The loss combines geometry transfer, energy expectation, spin penalty, and state-overlap penalties.
- Results: Reproduces the 8 lowest singlet and triplet states of the carbon dimer across 10 bond lengths (1.0–1.9 Å), matching semistochastic heat-bath CI (SHCI) benchmarks; ethylene torsion/pyramidalization scans agree with NES-VMC and MR-CI.
- Takeaway: One training run yields an entire geometrically transferable energy surface instead of point-by-point calculations.
More from Research
- ThinkJev: threshold-gated System 1/2 reasoning that stops thinking once confident — RobertTLange · 2026-11-07
- UCLA and Cedars-Sinai build a glioblastoma mouse model that resists immunotherapy like patients — anne_churchland · 2026-11-07
- Reorganization effect can reverse technology effect on AI-era wages, paper shows — soumitrashukla9 · 2026-11-07
- Meta's real-time avatars: distillation recipe cuts 120 model evals to 2, a 60x speedup for infinite video — altryne · 2026-11-07
- Free primer book derives math foundations from VAEs and diffusion to WGANs — kalyan_kpl · 2026-11-07
- Kenneth Stanley at ARC Prize summit: why open-endedness is essential to AGI but brutal to benchmark — GregKamradt · 2026-11-07