AI Paper Atlas

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

The Hard Positive Truth about Vision-Language Compositionality

Amita Kamath, Cheng-Yu Hsieh, Kai-Wei Chang, Ranjay Krishna

ECCV 2024 · Vision-language and multimodal › Human and embodied modeling · approach: Language-grounded scene text and structure · 9 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
finetune CLIP with hard negatives only, evaluated with hard positives addedCLIP ViT-B/32 · REPLACE benchmarkaccuracy drop (Orig -> Aug): avg 24.4 point drop vs CLIP's 14.9Hurt
hard negative finetuning (DAC-LLM) on REPLACECLIP ViT-B/32 · REPLACEBrittleness: 23.2 -> 40.1Hurt
add hard positives to hard negative finetuning (ablation)CLIP ViT-B/32 · REPLACEAugmented test accuracy: 55.7 -> 58.0Improved
add hard positives to hard negative finetuning (ablation)CLIP ViT-B/32 · REPLACEBrittleness: 21.0 -> 16.9Improved
add hard positives to hard negative finetuning (ablation)CLIP ViT-B/32 · SWAPBrittleness: 25.1 -> 22.9Improved
add generic paraphrase positives (SVLC+Pos) to SVLCSVLC · REPLACEAugmented test accuracy: 44.5 -> 45.0No clear effect
train with hard positives only (hard negative weight 0) (ablation)CLIP ViT-B/32 · REPLACEAugmented test accuracy: 57.6 (0.75 HN) -> 49.8 (0 HN)Hurt
increase hard negative loss weight (ablation)CLIP ViT-B/32 · REPLACEBrittleness: 15.8 -> 16.6 brittleness; aug acc 49.8 -> 57.6Mixed
finetune on both SWAP and REPLACE vs replace-only (ablation)CLIP ViT-B/32 · REPLACEAugmented test accuracy: 59.0 (replace-only) vs 58.0 (multi-task)No 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