AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Instruction-Oriented Preference Alignment for Enhancing Multi-Modal Comprehension Capability of MLLMs

Zitian Wang, Yue Liao, Kang Rong, Fengyun Rao, Yibo Yang, Si Liu

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Unified any-modality language models · 2 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
add contrastive sampling (perturbed-image responses) (ablation)IPA Round 1 · Ablation average (7 benchmarks)average score: 69.1 -> 70.1Improved
add contrastive sampling (ablation)IPA Round 1 · HallusionBenchscore: +2.9 on HallusionBenchImproved
add Round 2 self-improvement of generator and reviser (ablation)IPA · Ablation average (7 benchmarks)average score: 70.1 -> 70.4Improved
add Round 3 reference-guided feedback (ablation)IPA · Ablation average (7 benchmarks)average score: 70.4 -> 70.5Improved
use reference-guided sampling instead of reference-guided feedback in Round 3 (ablation)IPA · Ablation average (7 benchmarks)average score: 70.5 -> 69.8Hurt
add full IPA pipeline to Qwen2VL-7BQwen2VL-7B · HallusionBenchscore: 50.0 -> 54.3Improved
add full IPA pipeline to Qwen2VL-7BQwen2VL-7B · LLaVABenchscore: 76.6 -> 84.0Improved
add full IPA pipeline to Qwen2VL-7BQwen2VL-7B · MMHal-Benchscore: 3.6 -> 3.7No clear effect
apply IPA preference data to Qwen2.5VL-7BQwen2.5VL-7B · LLaVABenchscore: 75.2 -> 87.3Improved
apply IPA preference data to InternVL3-14BInternVL3-14B · LLaVABenchscore: 83.0 -> 89.9Improved
apply IPA preference data to Qwen2.5VL-72BQwen2.5VL-72B · MMStarscore: 71.6 -> 71.5No clear effect
align with IPA versus MMPRQwen2VL-7B · MMStarscore: 58.5 -> 61.7Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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