AI Paper Atlas

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

Cherry-Picking Gradients: Learning Low-Rank Embeddings of Visual Data via Differentiable Cross-Approximation

Mikhail Usvyatsov, Anastasia Makarova, Rafael Ballester-Ripoll, Maxim Rakhuba, Andreas Krause, Konrad Schindler

ICCV 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 2 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 CNN encoder before TT decomposition (ablation)C-Pic · OSIC (32x512x512)mLLL: -6.91 -> -6.73Improved
add CNN encoder before TT decomposition (ablation)C-Pic · BraTS 256^3RMSE (days): 83.9 -> 48.2Improved
use C-Pic low-rank TT bottleneck instead of 3D ResNet-34C-Pic · OSIC (32x512x512)mLLL: -6.86 -> -6.73Improved
use C-Pic instead of 3D ResNet-34C-Pic · BraTS 256^3RMSE (days): 48.7 -> 48.2No clear effect
raise encoding channel depth from 8 to 16 (ablation)C-Pic · BraTS 256^3RMSE (days): 51.1 -> 49.1Improved
raise encoding channel depth from 8 to 32 (ablation)C-Pic · BraTS 256^3RMSE (days): 51.1 -> 48.2Improved
raise TT rank r from 10 to 15 at n_dim=8 (ablation)C-Pic · BraTS 256^3RMSE (days): 51.1 -> 51.1No clear effect
double resolution per axis (8x voxels) with synthetic upsamplingC-Pic · OSICmLLL: -6.73 -> -6.73No clear effect
double resolution per axis with synthetic upsamplingC-Pic · BraTS (n_dim=8)RMSE (days): 51.1 -> 51.2No clear effect
raise input resolution to 64x1024x1024 for 3D ResNet-343D ResNet-34 · OSICforward/backward memory (GB): 7.0 -> 57.9Hurt
raise input resolution to 64x1024x1024 for C-PicC-Pic · OSICforward/backward memory (GB): 3.5 -> 4.2Mixed
C-Pic versus ResNet-34 training timeC-Pic · OSIC (32x512x512)training time (s/epoch): 4650 -> 51480Hurt

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