AI Paper Atlas

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

Joint Data-Task Generation for Auxiliary Learning

Hong Chen, Xin Wang, Yuwei Zhou, Yijian Qin, Chaoyu Guan, Wenwu Zhu

NeurIPS 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Data augmentation and generalization bias · 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 DTG-AuxL generated auxiliary data and task instead of STLResNet50 on CUB · CUBAcc(%): 80.46 -> 81.73Improved
Add manually collected auxiliary task via JTDS (unhelpful auxiliary)ResNet50 on CUB · CUBAcc(%): 80.46 -> 79.34Hurt
Add DTG-AuxL when the primary data is 5-shotResNet50 on CUB · CUB 5-shotAcc(%): 44.99 -> 52.33Improved
Add DTG-AuxL when the primary data is 10-shotResNet50 on CUB · CUB 10-shotAcc(%): 63.25 -> 67.68Improved
Switch backbone from ConvNet to ResNet18 so auxiliary task becomes beneficialmanual auxiliary learning · CIFAR100Acc(%): 50.35 -> 49.56Mixed
Remove label bias term from label generator (ablation)DTG-AuxL · CUBAcc(%): 79.28 -> 81.73Improved
Add label nonlinear term to generator (ablation)DTG-AuxL · Pet-CUBAcc(%): 66.58 -> 70.48Improved
Add feature nonlinear term to generator (ablation)DTG-AuxL · Pet-CUBAcc(%): 67.76 -> 70.48Improved
Add bi-level optimization of the generator (ablation)DTG-AuxL · Pet-CUBAcc(%): 68.53 -> 70.48Improved
Add instance regularization in upper update (ablation)DTG-AuxL · Pet-CUBAcc(%): 67.53 -> 70.48Improved
Add bi-level optimization of the generator (ablation)DTG-AuxL · ToysRMSE: 0.9163 -> 0.9153No clear effect
Replace learnable F-MixUp with joint data-task generator (ablation)Auxiliary data generation · CUB (ResNet18)Acc(%): 76.88 -> 77.75Improved

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