AI Paper Atlas

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

Trustworthy Multimodal Regression with Mixture of Normal-inverse Gamma Distributions

Huan Ma, Zongbo Han, Changqing Zhang, Huazhu Fu, Joey Tianyi Zhou, Qinghua Hu

NeurIPS 2021 · Vision-language and multimodal › Cross-modal learning · approach: Multimodal sentiment analysis · 16 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
fuse modality NIGs with NIG summation instead of simple average (ablation)MoNIG · SuperconductivityRMSE: 13.01 -> 12.19Improved
fuse modality NIGs with NIG summation instead of simple average (ablation)MoNIG · CT slicesRMSE: 2.87 -> 0.91Improved
use NIG summation instead of AU-weighted average (ablation)MoNIG · CT slicesRMSE: 5.71 -> 0.91Improved
use NIG summation instead of EU-weighted average (ablation)MoNIG · SuperconductivityRMSE: 13.03 -> 12.19Improved
add pseudo (feature-concatenation) modality (ablation)MoNIG · SuperconductivityRMSE: 12.19 -> 11.70Improved
add pseudo modality (ablation)MoNIG · CT slicesRMSE: 0.91 -> 0.79Improved
add pseudo modality (ablation)MoNIG · CMU-MOSEI unalignedMAE: 0.598 -> 0.612Hurt
add pseudo modality (ablation)MoNIG · CMU-MOSI word alignedAcc2: 80.2 -> 80.6Improved
modality-specific mixture of NIGs instead of evidential model on concatenated features under modality noiseevidential regression · Superconductivity eps=0.1RMSE: 44.80 -> 21.82Improved
MoNIG instead of evidential regression on concatenated features under modality noiseevidential regression · CT eps=0.1RMSE: 5.35 -> 2.23Improved
MoNIG instead of evidential regression for uncertainty qualityevidential regression · SuperconductivityUEIR (AU): 13.73 -> 12.65Improved
MoNIG instead of Gaussian/evidential baselines on sentimentmultimodal sentiment · CMU-MOSI word alignedMAE: 0.909 -> 0.951Mixed

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