AI Paper Atlas

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

Studying How to Efficiently and Effectively Guide Models with Explanations

Sukrut Rao, Moritz Böhle, Amin Parchami-Araghi, Bernt Schiele

ICCV 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Loss design and classifier theory · 18 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
use the Energy (EPG) loss instead of L1, PPCE or RRR* (ablation)model guidance with bounding boxes · PASCAL VOC 2007 / MS COCO 2014EPG vs F1 trade-offImproved
use the L1 loss instead of Energy, PPCE or RRR* (ablation)model guidance with bounding boxes · PASCAL VOC 2007IoU vs F1 trade-offImproved
use Energy loss rather than L1 to concentrate attributions on-object inside boxes (ablation)model guidance · VOC test subset with segmentation masksenergy within segmentation mask vs boxImproved
use B-cos explanations instead of Vanilla or X-DNN attributions at input layer (ablation)model guidance · PASCAL VOC 2007EPG @ F1: Vanilla 55.8 @ 69.0%, X-DNN 62.3 @ 68.9% -> B-cos 71.7 @ 79.4%Improved
add model guidance with bounding boxes to X-DNN (ablation)X-DNN ResNet-50 · PASCAL VOC 2007F1 and EPG: EPG +7.2 p.p., F1 +1.4 p.p.Improved
model guidance overall trade-off between localization and classificationmodel guidance · PASCAL VOC 2007 (i.i.d. test)F1Mixed
use only 1% of training images with box annotations (ablation)Energy loss guidance · PASCAL VOC 2007EPG: +23.0 p.p. EPG at -0.3 p.p. F1Improved
dilate bounding boxes by 10, 25 or 50% during training (ablation)Energy loss · PASCAL VOC 2007EPG and IoUNo clear effect
dilate bounding boxes by 10, 25 or 50% during training (ablation)L1 loss · PASCAL VOC 2007EPG and IoUHurt
apply guidance at deeper layer Mid2 instead of input (ablation)B-cos guidance · PASCAL VOC 2007input-level EPG / training time: 1.7x speed-up with similar localization gainsImproved
add model guidance with 1% bounding boxes under spurious correlation (ablation)B-cos model · Waterbirds-100worst-group / overall accuracy: Energy conventional: worst 56.1, overall 71.2Improved

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