AI Paper Atlas

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

Elodi: Ensemble Logit Difference Inhibition for Positive-Congruent Training

Yue Zhao, Yantao Shen, Yuanjun Xiong, Shuo Yang, Wei Xia, Zhuowen Tu, Bernt Schiele, Stefano Soatto

TPAMI 2024 · Label-efficient and robust learning › Recognition and long-tail · approach: Confidence calibration objectives · 7 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
distill from a homogeneous 8x ensemble with LDI loss (Elodi K=10) instead of no treatmentRN-18 to RN-50 update · ImageNetNFR (%): 4.30 -> 2.11Improved
use top-K=10 classes in LDI instead of all 1000 (ablation)Elodi · ImageNetNFR (%): 2.18 -> 2.11No clear effect
Elodi instead of targeted FD-KLRN-18 to RN-50 update · ImageNetNFR (%): 2.90 -> 2.11Improved
Elodi instead of dropout as implicit ensembleRN-18 to RN-50 update · ImageNetRel-NFR (%): 21.92 -> 13.23Improved
Elodi instead of Focal Distillation, data growth with more classesRN-18 to RN-50 · ImageNet (increasing classes)NFR (%): 5.45 -> 4.10Improved
Elodi instead of Focal Distillation, more samples per classRN-18 to RN-50 · ImageNet (increasing samples/class)NFR (%): 2.65 -> 2.21Improved
Elodi on AG News vs no treatmentBERT-base update · AG NewsNFR (%): 0.934 -> 0.289Improved
Elodi on chain of three models vs no treatmentRN-18, RN-50, RN-101 · ImageNetpairwise NFR (%): 4.28 -> 2.04Improved
add LDI loss toward old model when old model was not trained with Elodi (both updates)Elodi · ImageNet RN-18 to RN-50NFR (%): 2.98 -> 2.85Improved
use homogeneous ensemble instead of all-diff-weak heterogeneous ensemble (ablation)Elodi guiding ensemble · ImageNet RN-18 to RN-50NFR (%): 2.25 -> 2.18No clear effect
calibrate heterogeneous ensembles with temperature scaling (ablation)Elodi all-diff ensemble · ImageNet RN-18 to RN-50NFR (%): 2.25 -> 2.35No clear effect
guide new model with ensemble of new architecture instead of old RN-18 architecture (ablation)Elodi · ImageNet RN-18 to RN-50RN-50 error rate (%): 26.82 -> 23.15Improved

Later papers that compared with this method

1 later accepted paper reports ELODI as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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