AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

An Unsupervised Sampling Approach for Image-Sentence Matching Using Document-level Structural Information

Zejun Li, Zhongyu Wei, Zhihao Fan, Haijun Shan, Xuanjing Huang

AAAI 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Multilingual language representation pretraining · 4 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 intra-document negative sampling and dropout sub-document positive sampling, with the fine-grained Transformer modelMulLink · MSCOCOp@1: 95.0 -> 97.6Improved
Add intra-document negative sampling and dropout sub-document positive sampling, with the fine-grained Transformer modelMulLink · Story-DIIAUC: 82.9 -> 85.5Improved
Add intra-document negative sampling and dropout sub-document positive sampling, with the fine-grained Transformer modelMulLink · Story-SISAUC: 68.8 -> 70.2Improved
Add concepts to the Transformer (ablation)Ours · SISp@1: 75.8 -> 77.2Improved
Add Transformer to the backbone (ablation)Ours · SISAUC: 84.0 -> 85.5Improved
Add Transformer and concepts rather than neither (ablation)Ours · SISp@1: 74.6 -> 77.2Improved
Add dropout sub-document objective (D) (ablation)Ours with C+I objectives · SISAUC: 85.2 -> 85.5Improved
Add intra-document objective (I) (ablation)Ours with C+D objectives · SISAUC: 85.4 -> 85.5Improved
Add cross-document objective (C) (ablation)Ours with I+D objectives · SISAUC: 84.1 -> 85.5Improved
Add intra-document and dropout sub-document objectives (I+D) (ablation)Ours with C objective · SISAUC: 85.0 -> 85.5Improved
Add only concept features without Transformer modeling (ablation)Ours · SISAUC: 85.3 vs 85.5No clear effect
Replace Transformer with softmax pooling for objects and concepts (ablation)Ours · SISAUC: 85.5 -> 85.1Hurt

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