AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Compositional Attribute Imbalance in Vision Datasets

Yanbiao Ma, Jiayi Chen, Wei Dai, Dong Zhao, Zeyu Zhang, Yuting Yang, Bowei Liu, Jiaxuan Zhao, Andi Zhang

AAAI 2026 · Vision-language and multimodal › Human and embodied modeling · approach: Human-like visual cognition in VLMs · 1 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 CAS-based sampling weights to CutMixCutMix/FMix/SaliencyMix with ResNeXt50 · ImageNet-1ktop-1 accuracy (%): 49.78 -> 50.96Improved
add CAS-based sampling weights to SaliencyMixCutMix/FMix/SaliencyMix with ResNeXt50 · ImageNet-1ktop-1 accuracy (%): 47.35 -> 50.42Improved
add CAS-based sampling weights to CutMixCutMix/FMix/SaliencyMix with ResNeXt50 · ImageNet-1k (high-CAS subset)top-1 accuracy (%): 30.25 -> 33.79Improved
add CAS-based sampling weights to CutMixCutMix/FMix/SaliencyMix with ResNet18 · Stanford Carstop-1 accuracy (%): 82.49 -> 89.57Improved
add CAS-based sampling weights to FMixCutMix/FMix/SaliencyMix with ResNet18 · Stanford Carstop-1 accuracy (%): 75.72 -> 86.51Improved
add CAS-based sampling weights to CutMixCutMix/FMix/SaliencyMix with ResNeXt50 · Stanford Dogstop-1 accuracy (%): 69.14 -> 74.57Improved
add CAS-based sampling weights to CutMixCutMix/FMix/SaliencyMix with ResNet18 · CIFAR-100top-1 accuracy (%): 54.57 -> 55.28Improved
add CAS-based sampling weights (all augmentations)CutMix/FMix/SaliencyMix with ResNeXt50 · DTDtop-1 accuracy (%): +0.09 for FMix (99.56 -> 99.65)No clear effect
add CAS-sampling to CutMix on CLIP backbone vs DINOv2CutMix/FMix/SaliencyMix with DINOv2 · DTDoverall gain: +1.38 (CLIP) vs +0.97 (DINOv2)Mixed
set power parameter b to 1.2 (ablation)CutMix/FMix/SaliencyMix with ResNeXt50 · CIFAR-100 and ImageNettop-1 accuracy (%)Improved
apply CAS sampling on top of CutMixCutMix/FMix/SaliencyMix with ResNet18 · ImageNet-1Kmean CAS std: 36.8 -> 29.3Improved

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