AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

They Can See It but not Say It: Iterative Self Knowledge Re-expression in Visual Reasoning Relative Pose Identification

MENGYU WANG, Ken Deng, Yftah Ziser, Shay Cohen, Tiejun Ma

NeurIPS 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Spatial reasoning in VLMs · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Analyzes stable output biases of vision-language models on visual reasoning relative pose identification (VRRPI) and proposes an iterative self-knowledge re-expression approach to mitigate them.

Problem

Modern VLMs show weak multi-view 3D spatial reasoning on visual reasoning relative pose identification, where they disproportionately predict specific motion patterns regardless of visual evidence.

Approach

Identifies stable output biases in VLMs on VRRPI and proposes an iterative self-knowledge re-expression procedure to mitigate them; the abstract is truncated so the exact mechanism is not stated.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
apply SKR-ID (classification head + iterative median-balanced self-annotation)LLaVA-OneVision · VRRPI-BenchMacro-F1: 35.0 -> 58.0Improved
apply SKR-IDQwen3-VL · VRRPI-BenchMacro-F1: 39.5 -> 53.8Improved
apply SKR-IDVST-RL · VRRPI-BenchMacro-F1: 45.4 -> 53.7Improved
add object-centric debiasing prompt for self-annotation (ablation)LLaVA-OneVision · VRRPI-BenchMacro-F1: 35.0 -> 49.1Improved
add one round of SKR with classification head (ablation)LLaVA-OneVision · VRRPI-BenchMacro-F1: 49.1 -> 53.2Improved
add median-threshold balancing during iteration (ablation)LLaVA-OneVision · VRRPI-BenchMacro-F1: 54.0 -> 58.0Improved
remove object-centric prompt at initialization (ablation)LLaVA-OneVision · VRRPI-BenchMacro-F1: 58.0 -> 33.3Hurt
initialize with random balanced labels instead of prompted pseudo-labels (ablation)LLaVA-OneVision · VRRPI-BenchMacro-F1: 58.0 -> 50.8Hurt
use 0.7-quantile threshold instead of median (ablation)LLaVA-OneVision · VRRPI-BenchMacro-F1: 58.0 -> 47.2Hurt
train on 70/30 pseudo-label subset instead of balanced (ablation)LLaVA-OneVision · VRRPI-BenchMacro-F1: 58.0 -> 56.7Hurt
median thresholding on frozen VLM scores only (no head or fine-tuning) (ablation)LLaVA-OneVision · VRRPI-BenchMacro-F1: 35.0 -> 32.3 / 34.3 / 43.4Mixed
use self-consistency over 5 prompted responses (ablation)LLaVA-OneVision · VRRPI-BenchMacro-F1: 49.1 -> 49.8No clear effect

Surprising findings

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