AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

Understanding and Constructing Latent Modality Structures in Multi-Modal Representation Learning

Qian Jiang, Changyou Chen, Han Zhao, Liqun Chen, Qing Ping, Son Dinh Tran, Yi Xu, Belinda Zeng, Trishul Chilimbi

CVPR 2023 · Vision-language and multimodal › Cross-modal learning · approach: Multimodal fusion with missing modalities · 50 citations (OpenAlex)

Read the paper

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 deep feature separation loss (orthogonal + contrastive + uniformity on independent features)CLIP two-tower · Zero-shot CIFAR10/CIFAR100/ImageNet1K and linear probingTop1 accuracyMixed
add Brownian-bridge lossCLIP two-tower · CIFAR10 and CIFAR100 zero-shotTop1 accuracyHurt
add Brownian-bridge lossCLIP two-tower · ImageNet1K zero-shot and linear probingTop1 accuracy: +3.73 zero-shot top1 vs CLIP; +1.79 linear average vs CLIPImproved
add geometric consistency lossCLIP two-tower · CIFAR10 and CIFAR100 zero-shotTop1 accuracy: CIFAR10 +11.41; CIFAR100 +7.65 top1 vs CLIPImproved
add geometric consistency lossCLIP two-tower · Natural distribution shifts (ImageNetV2, Sketch, A, R)Top1 accuracyImproved
add deep feature separation, Brownian-bridge, or geometric consistency loss individuallyCLIP two-tower · Linear probing average over 14 benchmarksTop1 accuracy: Sep +5.68, Br +1.79, GC +2.96 vs CLIPImproved
add Brownian-bridge lossALBEF fusion model · VQA test-devaccuracy: +0.88 vs ALBEFImproved
add deep feature separation lossALBEF fusion model · NLVR2 devaccuracy: +0.69 vs ALBEFImproved
add all three regularizers together (ablation)ALBEF fusion model · VQA test-dev, NLVR2 dev, SNLI-VE testaccuracy: VQA +0.74 vs ALBEF; NLVR2 dev +1.82 vs ALBEF; SNLI-VE test +0.32 vs ALBEFImproved
add deep feature separation, Brownian-bridge, or geometric consistency loss individually (ablation)ALBEF fusion model · Flickr30K zero-shot text retrieval R@1 (COCO pretrain)R@1: Sep +7.6, Br +7.0, GC +5.9 vs ALBEFImproved
add Brownian-bridge loss (ablation)ALBEF fusion model · Flickr30K zero-shot image retrieval R@1 (COCO pretrain)R@1: +6.3 vs ALBEF; +1.8 vs CODISImproved
add modality-gap-reducing alignment loss (1/<Z_T,Z_V>^2) (ablation)ALBEF-style pretraining on COCO · Zero-shot Flickr30K retrievalretrieval performanceNo clear effect

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