AI Paper Atlas

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

Uncertainty-Aware Hierarchical Refinement for Incremental Implicitly-Refined Classification

Jian Yang, Kai Zhu, Kecheng Zheng, Yang Cao

NeurIPS 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Learning from unlabeled and positive-unlabeled data · 1 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 hierarchical distribution alignment (HDA) to the baseline (ablation)UAHR baseline · IIRC-CIFAR config 0, phase 21PW-JS: 17.78 -> 18.38Improved
add global representation extension (GRE) to the baseline (ablation)UAHR baseline · IIRC-CIFAR config 0, phase 21PW-JS: 17.78 -> 18.32Improved
add both GRE and HDA (ablation)UAHR baseline · IIRC-CIFAR config 0, phase 21PW-JS: 17.78 -> 19.05Improved
add HDA to the baseline at phase 5 (ablation)UAHR baseline · IIRC-CIFAR config 0, phase 5PW-JS: 26.48 -> 29.31Improved
add GRE on top of HDA (ablation)UAHR · IIRC-CIFAR config 0, phase 21PW-JS: 18.38 -> 19.05Improved
add GRE or HDA at phase 0 (initial phase) (ablation)UAHR · IIRC-CIFAR config 0, phase 0PW-JS: 78.35 -> 77.53Hurt
use hierarchical uncertainty (entropy output + feature std) instead of common entropy uncertainty (ablation)UAHR · IIRC-CIFAR, phase 5PW-JS: 25.70 -> 30.11Improved
use hierarchical uncertainty instead of common uncertainty (ablation)UAHR · IIRC-CIFAR, phase 10PW-JS: 20.89 -> 25.31Improved
use hierarchical uncertainty instead of common uncertainty (ablation)UAHR · IIRC-CIFAR, phase 15PW-JS: 19.55 -> 23.56Improved
use hierarchical uncertainty instead of common uncertainty (ablation)UAHR · IIRC-CIFAR, phase 21PW-JS: 15.32 -> 19.05Improved

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