InvDesFlow-AL
A diffusion model generates crystals, a committee of predictors and relaxation checks picks the best, and the model is fine-tuned on them.
The loop
- Proposes
Crystal structures from a diffusion model (EGNN backbone) pretrained on Alex-MP-20 and GNoME crystals
- Judged by
Relaxation with DPA-2, then a committee score on formation energy and Ehull or on superconductor Tc
- Steers
The top-scored generated crystals become the next fine-tuning set, repeated over multiple rounds
Headline result
Generated average formation energy fell over five iterations (means -1.14, -2.03, -2.93, -3.56, -3.77; units not stated in the text). The authors report 1,598,551 materials with Ehull below 50 meV, and Li2AuH6 as a conventional superconductor with a transition temperature of about 140 K from first-principles electron-phonon calculations. Crystal structure prediction RMSE is 0.0423 angstrom, reported as a 32.96% improvement over existing generative models.
What this does not prove
The paper gives two counts for low-Ehull materials (1,598,551 in text, 1,610,600 in a figure caption) and does not say how Ehull is queried. The 140 K value is a calculation, not a measurement, and not every evaluation is DFT.
How it generalizes the harness
A diffusion model is fine-tuned over rounds on crystals chosen by relaxation and learned predictors, with many crystals generated per round; DFT is used explicitly only for the superconductor task.
Classification
| What it proposes | Crystal structure |
|---|---|
| Generative model | Diffusion model on atom types, lattice and fractional coordinates (EGNN) |
| Algorithmic authority | DPA-2 relaxation, committee scoring, Bohrium Ehull query, DFT for the superconductor task |
| Feedback | Checks at several fidelities |
| What gets updated | Model weights and training data |
| When it runs | During training |
| Validity | Approximate physics (simulation, not lab results) |
| Evidence | Experiments reported |
| Lineage | Successor |