AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

When and How Human Curation Backfires: Preference Alignment under Multi-Model Self-Consuming Loop

Yang Zhang, Xiukun Wei, Xueru Zhang

ICML 2026 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Self-supervision in supervised pipelines · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

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
increase human-curated fraction for model theta while its data is heavily cross-modeltwo-model self-consuming loop (theta) · CIFAR-10 diffusion, A4Jp(theta*) rewardHurt
increase human curation for theta when cross-model data fraction is small (<=0.3)two-model self-consuming loop (theta) · CIFAR-10 diffusion, A1-A3 and A5Jp(theta*) rewardImproved
increase curation data for model phi when its training data comes entirely from conflicting model thetatwo-model self-consuming loop (phi) · CIFAR-10 diffusionJq(phi*) rewardHurt
reduce influence of theta on phi (lower cross-model fraction) while curating thetatwo-model self-consuming loop (phi) · CIFAR-10 diffusion, A4Jq(phi*) rewardImproved
strong coupling between models (A6)two-model self-consuming loop · CIFAR-10 diffusion, A6rewardsHurt
increase coupling scale t with fixed curation lambda_H=0.4Gaussian coupled system · Gaussian models, t in [0.05,1]Jq(phi*) derivativeHurt
train only on cross-model data (preference domain mismatch)Qwen2.5-0.5B summarization/paraphrase models · Qwen2.5-0.5B, lambda cross=1target-domain rewardNo clear effect
increase sample size n used for model updatesGaussian experiment estimators · Gaussian models, n in [4,64]statistical error vs theoryImproved

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