AI Paper Atlas

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

Constructing Superior Representations Beyond the Original Documents via a Contrastive Gaussian Fusion Network for Clustering

Ao Shen, Ruizhang Huang, Jingjing Xue, Ruina Bai

AAAI 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Contrastive representation theory · 0 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 consistency enhancement (beta<1 KL weight) in place of standard VAE regularization (ablation)CGFN without clustering layer · BBCNMI: 51.20 -> 63.45Improved
add consistency enhancement in place of standard VAE regularization (ablation)CGFN without clustering layer · Reuters-10kNMI: 43.28 -> 54.09Improved
add contrastive Gaussian fusion with neighbor data to single-input VAE (ablation)CGFN-f · AbstractNMI: 33.02 -> 45.75Improved
add contrastive Gaussian fusion with neighbor data to single-input VAE (ablation)CGFN-f · BBCSportsNMI: 34.05 -> 50.91Improved
add clustering layer (self-training KL objective) (ablation)CGFN-c · BBCSportsNMI: 58.40 -> 85.79Improved
use CGFN instead of best baselinedocument clustering · Reuters-10kACC: 77.15 -> 75.71Hurt
use CGFN instead of best baselinedocument clustering · four datasetsNMI: +11.73, +2.81, +17.50, +3.50 points on BBC, Abstract, BBCSports, Reuters-10kImproved

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