AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

ART: Adaptive Relation Tuning for Generalized Relation Prediction

Gopika Sudhakaran, Hikaru Shindo, Patrick Schramowski, Simone Schaub-Meyer, Kristian Kersting, Stefan Roth

ICCV 2025 · Vision-language and multimodal › CLIP and zero-shot · approach: Test-time adaptation of VLMs · 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
tune on adaptively selected 12% of data (ART) instead of balanced random samplingInstructBLIP Vicuna-7B · GQAmR@20: 13.2 -> 18.9Improved
use ART instead of random sampling on RecallInstructBLIP Vicuna-7B · GQAR@20: 59.0 -> 40.1Hurt
tune on random 12% subset instead of full VG train data (ablation)InstructBLIP Vicuna-7B · VGmR@20: 14.1 -> 19.6Improved
use balanced random sampling instead of random (ablation)InstructBLIP Vicuna-7B · VGmR@20: 19.6 -> 42.4Improved
add adaptive sampling (hFN, tFP, tFN, hTP) on top of balanced sampling (ablation)ART · VGmR@20: 42.4 -> 46.4Improved
select by TP/FN/FP-specific criteria instead of entropy-only (ablation)ART · VGmR@20: 44.6 -> 46.4Improved
add high-level relation types to the prompt (ablation)ART instruction set · VGmR@20: 37.5 -> 39.4Improved
replace random negatives with one counter-negative (ablation)ART instruction set · VGmR@20: 40.9 -> 46.4Improved
increase counter-negatives per sample from 1 to 2 (ablation)ART instruction set · VGmR@20: 46.4 -> 40.3Hurt
use adaptive thresholds instead of fixed thresholds (ablation)ART · VGmR@20: 44.8 -> 46.4 (best fixed vs adaptive)Improved
use ART-predicted scene graphs in DeiSAM instead of ground-truth graphsDeiSAM · Deictic segmentation VG (w/o bbox)mAP: 4.77 -> 25.07Improved
ART on detected RAM boxes vs random samplingInstructBLIP Vicuna-7B · OI-v4 with RAM boxesmR@20: 6.7 -> 24.0Improved

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