Skip to content

Perov-5

18,928 cubic ABX3 perovskites with relaxed structures, formation enthalpy (heat_all, eV/atom) and direct band gap (dir_gap, eV). The CDVAE split of the Castelli et al. (2012) perovskite set; the dataset behind the published MEIDNet model and the live Studio.

18,928 rows · split: CDVAE: 11,356 train / 3,787 val / 3,785 test · properties: heat_all, dir_gap · role in the paper: the published multimodal benchmark: alignment, property reconstruction and inverse design · dataset source

5results
2reproduced here
3categories
DFT validatedhighest evidence

Read a row

Rows are compared only within this dataset and category; click a column header to sort. The ladder ●○○○○ … ●●●●● is the evidence level (generated → ML filtered → MLIP validated → DFT validated → experimentally validated). Published = taken from the paper; MEIDNet verified = reproduced here from the submitted configuration.

Representation quality

result split inputs cosine_matched l2_matched retrieval_top1 retrieval_top5 mae_heat_all r2_heat_all mae_dir_gap r2_dir_gap n_evaluated evidence status
MEIDNet (paper): alignment of structure and property la… validation structure + 2 properties 0.97 0.24 – – – – – – – ●○○○○ generated Published (from the paper)
Shipped checkpoint re-evaluated on the training split (… train structure + 2 properties 0.771 0.676 0.0101 0.0405 0.397 eV/atom 0.486 2.78 eV -29.8 11,356 materials ●○○○○ generated MEIDNet verified
Shipped checkpoint re-evaluated on the held-out validat… val structure + 2 properties 0.771 0.676 0.0253 0.0908 0.392 eV/atom 0.497 2.79 eV -35.8 3,787 materials ●○○○○ generated MEIDNet verified
How to read the representation quality table
  • cosine_matched — mean cosine similarity between the structure and property latents of the same material (1 = aligned)
  • l2_matched — mean L2 distance between those two latents (0 = identical) (lower is better)
  • retrieval_top1 — fraction of validation materials whose property latent is nearest to its own structure latent
  • retrieval_top5 — the same within the five nearest
  • mae_heat_all — mean absolute error of heat_all predicted from the structure alone (physical units) (lower is better)
  • r2_heat_all — coefficient of determination of that prediction; 0 = no better than the mean, 1 = perfect
  • mae_dir_gap — mean absolute error of dir_gap predicted from the structure alone (physical units) (lower is better)
  • r2_dir_gap — coefficient of determination of that prediction; 0 = no better than the mean, 1 = perfect
  • n_evaluated — materials in the evaluation split

Representation rows describe how well the shared latent space holds structures and properties; they say nothing about whether a generated material is real.

cosine_matched (higher is better)MEIDNet (paper): alignment of structur0.97Shipped checkpoint re-evaluated on the0.771Shipped checkpoint re-evaluated on the0.771
retrieval_top1 (higher is better)Shipped checkpoint re-evaluated on the0.0253Shipped checkpoint re-evaluated on the0.0101
mae_heat_all (lower is better)Shipped checkpoint re-evaluated on the0.397Shipped checkpoint re-evaluated on the0.392
mae_dir_gap (lower is better)Shipped checkpoint re-evaluated on the2.79Shipped checkpoint re-evaluated on the2.78
dir_gap: predicted from the structure vs. true (3,787 validation materials)02460246true dir_gap (eV)predicted dir_gap
heat_all: predicted from the structure vs. true (3,787 validation materials)024024true heat_all (eV/atom)predicted heat_all
cosine similarity of the two latents, per material0.650.70.750.80.850200400600800cosine(structure latent, property latent)count0.636–0.642: 10.672–0.678: 10.689–0.695: 10.695–0.701: 10.701–0.707: 20.707–0.713: 30.713–0.719: 30.719–0.725: 60.725–0.731: 50.731–0.737: 160.737–0.743: 280.743–0.748: 520.748–0.754: 1080.754–0.76: 2970.76–0.766: 7360.766–0.772: 7540.772–0.778: 5820.778–0.784: 7710.784–0.79: 2510.79–0.796: 640.796–0.802: 290.802–0.808: 290.808–0.813: 180.813–0.819: 170.819–0.825: 50.825–0.831: 30.831–0.837: 30.855–0.861: 1mean

Conditional design

result split inputs n_generated n_sun sun_rate evidence status
MEIDNet (paper): inverse design of perovskites from pro… – structure + 2 properties 140 structures 19 structures 0.136 ●●●●○ DFT validated Published (from the paper)
How to read the conditional design table
  • n_generated — candidates the search produced
  • n_sun — stable, unique and novel candidates
  • sun_rate — n_sun / n_generated

Design rows are funnels: every number after n_generated is a subset of the one before. A high SUN rate on few candidates is weaker evidence than a lower rate on many.

MEIDNet (paper): inverse design of perovskites from property targets: from generated to validatedgenerated140stable, unique, novel19

Validation level

result split inputs n_screened evidence status
MLIP screening of the shipped paper candidates (MACE-MP… – structure + 2 properties 26 structures ●●●○○ MLIP validated Community submitted
How to read the validation level table
  • n_screened — structures relaxed with the MLIP

Validation rows count structures that survived each check. n_stable depends on the stability threshold used (0.10 eV/atom above the hull by default).

Insights

  • Best cosine_matched: MEIDNet (paper): alignment of structure and property latents on Perov-5 (0.97), published (from the paper).
  • Best l2_matched: MEIDNet (paper): alignment of structure and property latents on Perov-5 (0.24), published (from the paper).
  • Best retrieval_top1: Shipped checkpoint re-evaluated on the held-out validation split (this code) (0.0253), meidnet verified.
  • Best retrieval_top5: Shipped checkpoint re-evaluated on the held-out validation split (this code) (0.0908), meidnet verified.
  • Best mae_heat_all: Shipped checkpoint re-evaluated on the held-out validation split (this code) (0.392 eV/atom), meidnet verified.
  • Best r2_heat_all: Shipped checkpoint re-evaluated on the held-out validation split (this code) (0.497), meidnet verified.
  • Best mae_dir_gap: Shipped checkpoint re-evaluated on the training split (this code) (2.78 eV), meidnet verified.
  • Best r2_dir_gap: Shipped checkpoint re-evaluated on the training split (this code) (-29.8), meidnet verified.
  • Evidence levels present: generated, MLIP validated, DFT validated. 2 of 5 rows have been reproduced here.
  • 2 row(s) are quoted from the paper; a verified re-run of the same checkpoint, where it exists, is the row to trust for the exact numbers of this code version.
  • A re-run of MEIDNet (paper): alignment of structure and property latents on Perov-5 here did not agree (cosine_matched 0.97 → 0.771; l2_matched 0.24 → 0.676). The row keeps its status; see its page for the full comparison.