AI Paper Atlas

Research map / Large language models / Alignment and preferences

Iterative Perceptual Alignment for VLMs via Deterministic Reconstruction Feedback

Xiaorui Chen, Hanzhong Guo, Nizhe Cai, Jieliang Luo

ECCV 2026 · Large language models › Alignment and preferences · approach: Robust and calibrated DPO · 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
DPO on reconstruction-feedback preference data (>0.03, 5K)Qwen3-VL-2B-Instruct · Hallusionaccuracy (%): 46.27 -> 57.83Improved
DPO on reconstruction-feedback dataInternVL3.5-2B-HF · POPEaccuracy (%): 86.67 -> 88.66Improved
DPO on reconstruction-feedback dataInternVL3.5-1B-HF · Hallusionaccuracy (%): 54.47 -> 51.10Hurt
DPO on reconstruction-feedback dataInternVL3.5-1B-HF · MM-Benchaccuracy (0-1): 0.32 -> 0.49Improved
DPO on reconstruction-feedback dataInternVL3.5-2B-HF · RealWorldQAaccuracy (0-1): 0.39 -> 0.35Hurt
use 3K auto data (>0.05) instead of 5K (>0.03) (ablation)DPO on Qwen3-VL-2B · MM-Benchaccuracy (0-1): 0.67 -> 0.70Improved
use 5K auto data (>0.03) instead of 3K (>0.05) (ablation)DPO on Qwen3-VL-2B · Hallusionaccuracy (%): 56.99 -> 57.83Improved
replace human-annotated RLHF-V 3.5k with automatic 3K dataDPO on InternVL3.5-2B · POPEaccuracy (%): 88.56 -> 88.66No clear effect
train reward model on reconstruction-based pairs (about 20% of data)T2I-Consistency-Score · zero-shot RLHF-V / VLFeedbackaccuracy (%): 74.12 vs 65.25 (RLHF-V); 67.43 vs 77.41 (POVID)Mixed
add score decomposition with text-only branchreward model · in-domain preferenceaccuracyNo clear effect

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