AI Paper Atlas

Research map / Label-efficient and robust learning / Noisy labels

Teaching with Soft Label Smoothing for Mitigating Noisy Labels in Facial Expressions

Tohar Lukov, Na Zhao, Gim Hee Lee, Ser-Nam Lim

ECCV 2022 · Label-efficient and robust learning › Noisy labels · approach: Instance-dependent noise modeling · 44 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 SLS over teacher logits to mean-teacher (ablation)MT+U · RAF-DB 30% symmetric noiseaccuracy: 85.13 -> 86.94Improved
add mean-teacher to SLS (ablation)SLS vanilla · RAF-DB 30% symmetric noiseaccuracy: 85.33 -> 86.82Improved
add uncertainty module to MT+SLS (ablation)MT+SLS · RAF-DB 30% symmetric noiseaccuracy: 86.82 -> 86.94Improved
use instance-aware SLS instead of LSR* (ablation)teacher smoothing · RAF-DB 30% symmetric noiseaccuracy: 85.85 -> 86.94Improved
use non-zero low confidences (epsilon>0) in SLS (ablation)SLS · RAF-DB 30% symmetric noiseaccuracy: 85.30 -> 86.94Improved
apply SLS to teacher logits instead of student logits (ablation)SLS · RAF-DB 30% symmetric noiseaccuracy: 85.23 -> 86.94Improved
apply SLS to both student and teacher logits instead of teacher only (ablation)SLS · RAF-DB 30% symmetric noiseaccuracy: 86.94 -> 85.36Hurt
add SLS to MT in noisy-part analysis (ablation)MT+U · RAF-DB 30% symmetric noisenoisy samples predicted with injected wrong label: 28.33 -> 13.57Improved
increase epsilon beyond 0.3 (ablation)SLS · RAF-DB 30% symmetric noiseaccuracyHurt
raise tau toward 0.4 so only one high-confidence class remains (ablation)SLS · RAF-DB 30% symmetric noiseaccuracyHurt
use SOFT instead of DMUE under symmetric noiseSOFT · RAF-DB symmetric noise 10-70%accuracy: 5.7%, 7%, 7.5% and 26.5% on RAF-DB with 10%, 20%, 30% and 70% noise ratesImproved
use SOFT instead of DMUE under symmetric noiseSOFT · AffectNet-8 symmetric noiseaccuracy: 0.1%, 1.9%, 2.8% on AffectNet-8 with 10%, 20%, 30% noise ratesImproved

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