AI Paper Atlas

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

QuMAB: Query-based Multi-annotator Behavior Pattern Learning

Liyun Zhang, Zheng Lian, Hong Liu, Takanori Takebe, Shozo Nishii, Yuta Nakashima

AAAI 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Learning from unlabeled and positive-unlabeled data · 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 Q-Former with annotator queries over frozen encoder (ablation)Base encoder + classifiers · AMERaccuracy: 0.28 -> 0.84Improved
use individual classifiers instead of a unified classifier (ablation)QuMAB · AMERaccuracy: 0.61 -> 0.84Improved
use individual classifiers instead of a unified classifier (ablation)QuMAB · STREET Happinessaccuracy: 0.50 -> 0.63Improved
add shared self-attention among annotator queries (ablation)QuMAB · AMERaccuracy: 0.73 -> 0.84Improved
add shared self-attention among annotator queries (ablation)QuMAB · STREET Safetyaccuracy: 0.46 -> 0.58No clear effect
model each annotator then aggregate (Post-mv) instead of pooling before modeling (ablation)QuMAB · AMER consensusaccuracy: 0.43 -> 0.60Improved
use QuMAB instead of PADL under 40% removed annotationsPADL · AMERaccuracy: 0.58 -> 0.66Improved
use QuMAB instead of PADL under 40% removed annotationsPADL · STREET Happinessaccuracy: 0.43 -> 0.52Improved
use QuMAB instead of PADL for average drop under sparsityPADL · STREET and AMERaccuracy drop: 27.4% -> 20.4%Improved
use QuMAB instead of baselines for consensus predictionMaDL · AMERaccuracy: 0.57 -> 0.60Improved
use QuMAB instead of baselines for consensus predictionPADL · STREET Healthinessaccuracy: 0.50 -> 0.58Improved
use QuMAB instead of baselines for individual annotator modelingMaDL · STREET Safetyaccuracy: 0.46 -> 0.58Improved

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