Quartz 4

Home

❯

tags

❯

Inference

❯

Ensembling

Ensembling

Created Sep 26, 2026Modified Sep 07, 2026

Combining predictions, scores, prompts, or samples from multiple model runs. Weight merging is documented separately under training/fine-tuning.

6 items with this tag.

  • Aug 25, 2026

    Ensembling structure prediction methods for filtering improves recovery

    • evidence/design-validation
    • design/binders
    • inference/ensembling
  • Aug 25, 2026

    Zero-shot protein fitness prediction using sequence-only NNs can be improved by averaging predictions from many orthologs

    • prediction/variant-effects
    • inference/ensembling
  • Jul 28, 2026

    AF3 binding affinity predictions are orthogonal to those made by force fields and other neural networks

    • prediction/binding
    • inference/ensembling
  • Apr 21, 2026

    Averaging logits from multiple sources can improve fitness prediction

    • prediction/variant-effects
    • prediction/stability-expression
    • inference/ensembling
  • Apr 21, 2026

    Taking the minimum logit can outperform averaging logits when ensembling models for pathogenicity prediction

    • prediction/variant-effects
    • prediction/confidence
    • inference/ensembling
  • Apr 21, 2026

    Variant effect prediction with homology-aware PLMs improves with ensembling of multiple prompts

    • prediction/variant-effects
    • inference/conditioning
    • inference/ensembling

Created with Quartz v4.5.2 © 2026

  • GitHub
  • Discord Community