AI Paper Atlas

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

Task Prototype-Based Knowledge Retrieval for Multi-Task Learning from Partially Annotated Data

Youngmin Oh, Hyung-Il Kim, Jung Uk Kim

AAAI 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Dense features and foundation priors · 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 vector quantization with TAE loss (ablation)prototype-based knowledge retrieval MTL · PASCAL-Context one-labelSemseg mIoU: 44.73 -> 44.83No clear effect
add task knowledge embedding (TKE) loss (ablation)prototype-based knowledge retrieval MTL · PASCAL-Context one-labelSemseg mIoU: 44.83 -> 58.21Improved
add task consistency (TC) loss (ablation)prototype-based knowledge retrieval MTL · PASCAL-Context one-labelSemseg mIoU: 58.21 -> 59.78Improved
add TKE loss (ablation)prototype-based knowledge retrieval MTL · PASCAL-Context one-labelNormal mErr: 16.22 -> 15.67Improved
increase task prototype slot dimension from 1024 to 2048 (ablation)prototype-based knowledge retrieval MTL · NYUD-v2 one-labelSemseg mIoU: 45.95 -> 45.33Hurt
reduce slot dimension from 1024 to 256 (ablation)prototype-based knowledge retrieval MTL · NYUD-v2 one-labelSemseg mIoU: 45.95 -> 44.91Hurt
use proposed method instead of DiffusionMTL (Feature)prototype-based knowledge retrieval MTL · PASCAL-Context one-labelSemseg mIoU: 57.78 -> 59.78Improved
use proposed method instead of DiffusionMTL (Feature)prototype-based knowledge retrieval MTL · PASCAL-Context one-labelBoundary odsF: 64.50 -> 65.10Improved
use proposed method instead of TaskPrompterprototype-based knowledge retrieval MTL · NYUD-v2 one-label ViT-LSemseg mIoU: 48.68 -> 50.08Improved
use proposed method instead of DiffusionMTL (Prediction)prototype-based knowledge retrieval MTL · PASCAL-Context ResNet-50Semseg mIoU: 60.92 -> 62.23Improved
proposed method vs single-task learningprototype-based knowledge retrieval MTL · NYUD-v2 random-labelNormal mErr: 24.65 vs 24.67Mixed

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