AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Constrained Diffusion for Protein Design with Hard Structural Constraints

Jacob K Christopher, Austin Seamann, Jingyi Cui, Sagar D Khare, Ferdinando Fioretto

ICLR 2026 · Image generation › Autoregressive and latent generation · approach: Diffusion for protein and antibody design · 4 citations (OpenAlex)

Read the paper

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use ADMM-based constrained diffusion instead of RFDiffusion baselinesRFDiffusion · PDZ domainUsable percentage (%): 0.0 -> 21.0Improved
use ADMM-based constrained diffusion instead of best baseline (Recenter + CGD)RFDiffusion · Molecule encapsulationUsable percentage (%): 24.2 -> 97.8Improved
use ADMM-based constrained diffusion instead of Recenter + CGDRFDiffusion · Molecule encapsulationConstraint satisfaction (%): 27.4 -> 100.0Improved
add constraint guidance (CGD) to standard RFDiffusionRFDiffusion · Molecule encapsulationConstraint satisfaction (%): 0.0 -> 21.6Improved
add recentering to CGDRFDiffusion CGD · Molecule encapsulationConstraint satisfaction (%): 21.6 -> 27.4Improved
add constrained diffusion over CGD baselineCGD · Molecule encapsulationRadius of gyration (A): 23.9 -> 14.8Improved
use our method instead of RFDiffusion StandardRFDiffusion · PDZ domainSecondary structure (%): 80.8 -> 85.0Improved
use our method instead of CGD on PDZCGD · PDZ domainStructure realism (%): 38.2 -> 21.0Hurt
use ADMM with local+global constraints in projected diffusion (ablation)Projected Diffusion Model · PDZ domainConstraint satisfaction (%): 100.0 (Global only) -> 0.0 (Failed to Converge)Hurt
replace projection (Global PDM) with our methodProjected Diffusion Model Global · PDZ domainStructure realism (%): 0.0 -> 97.8Improved
increase constraint weight schedule from 10/t to 100/t (ablation)Ours · PDZ 6ubhUsable percentage (%): 80.0 -> 100.0Improved
use lambda=inf instead of 100/t (ablation)Ours · PDZ 6ubhUsable percentage (%): 100.0 -> 90.0Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy