AI Paper Atlas

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

Understanding Synonymous Referring Expressions via Contrastive Features

Yiwen Chen, Yi–Hsuan Tsai, Ming–Hsuan Yang

IJCV 2022 · Vision-language and multimodal › Human and embodied modeling · approach: Referring expression comprehension · 6 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 synonymous-sentence contrastive feature lossesOurs (LSTM) · RefCOCO valaccuracy: 76.63 -> 79.09Improved
add synonymous-sentence contrastive feature lossesOurs (BERT) · RefCOCO valaccuracy: 77.02 -> 79.63Improved
add losses when jointly training on three datasetsOurs (BERT) · RefCOCO val jointaccuracy: 75.17 -> 82.42Improved
add image-level feature learning only (ablation)baseline · RefCOCO val jointaccuracy: 75.17 -> 77.82Improved
add instance-level contrastive loss only (ablation)baseline · RefCOCO val jointaccuracy: 75.17 -> 79.33Improved
use contrastive instead of triplet loss at instance level (ablation)final model · RefCOCO val jointaccuracy: 80.61 -> 82.42Improved
add negative mining (ablation)final model · RefCOCO val jointaccuracy: 82.05 -> 82.42Improved
add GCN (ablation)final model · RefCOCO val jointaccuracy: 81.40 -> 82.42Improved
train with losses on RefCOCOg, evaluate unseen on Ref-Reasoning one-objectOurs · Ref-Reasoning (one object)accuracy: 50.26 -> 53.49Improved
plain pre-training on RefCOCO datasets without proposed loss (ablation)Ours (baseline) · Ref-Reasoning one-objectaccuracy: 76.43 -> 76.54No clear effect
add loss in both pre-training and fine-tuning (ablation)Ours · Ref-Reasoning one-objectaccuracy: 76.54 -> 81.94Improved
add loss in both pre-training and fine-tuning (ablation)Ours · ReferItGame testaccuracy: 57.13 -> 61.71Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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