AI Paper Atlas

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

Left–Right Symmetry Breaking in CLIP-Style Vision-Language Models Trained on Synthetic Spatial-Relation Data

Takaki Yamamoto, Chihiro Noguchi, Toshihiro Tanizawa

ICML 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Recaptioning and text-supervised pretraining · 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
increase number of training object labels (label diversity) (ablation)CLIP-style Transformer on 1D toy data · unseen-pair, seen-pair and single-object generalizationimage-to-text retrieval accuracyImproved
increase layout diversity (spatial configurations) (ablation)CLIP-style Transformer on 1D toy data · generalization accuracyimage-to-text retrieval accuracyNo clear effect
add weight decay regularization (ablation)CLIP-style training · toy generalizationgeneralization accuracyImproved
zero the token-position (EP) cross term in attention logits (ablation)vision encoder · unseen-pair generalizationleft-right discrimination accuracy: drops to near 0.5Hurt
ablate position-dependent value term PW_V (ablation)vision encoder · unseen-pair generalizationaccuracy: decreases to 0.5Hurt
train with both left and right textual captions using 1-layer text encoder (ablation)CLIP-style model · unseen-pair generalizationtop-2 retrieval correctnessHurt
add a second layer to the text encoder with both left and right captions (ablation)text encoder · unseen-pair generalizationgeneralization accuracyImproved
extend from 1D to 4x4 2D imagestoy model · unseen-pair generalizationaccuracy: 98% accuracyNo 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