AI Paper Atlas

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

Learning from Label Proportions with Generative Adversarial Networks

Jiabin Liu, Bo Wang, Zhiquan Qi, YingJie Tian, Yong Shi

NeurIPS 2019 · Label-efficient and robust learning › Recognition and long-tail · approach: Loss design and classifier theory · 6 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
Replace DLLP with LLP-GANLLP learning · MNIST, bag sizes 16, 32, 64, and 128test error (%): 1.23 (0.100) -> 1.10 (0.026); 1.33 (0.094) -> 1.23 (0.088); 1.57 (0.088) -> 1.40 (0.089); 3.55 (0.27) -> 3.49 (0.27)Improved
Replace DLLP with LLP-GANLLP learning · SVHN, bag sizes 16, 32, 64, and 128test error (%): 4.45 (0.069) -> 4.03 (0.021); 5.29 (0.54) -> 4.83 (0.51); 5.80 (0.91) -> 5.42 (0.59); 39.73 (1.60) -> 11.17 (1.12)Improved
Replace DLLP with LLP-GANLLP learning · CIFAR-10, bag sizes 16, 32, 64, and 128test error (%): 19.70 (0.77) -> 13.68 (0.35); 34.39 (0.82) -> 16.23 (0.43); 68.32 (1.34) -> 21.03 (1.82); 82.89 (2.66) -> 27.39 (4.31)Improved
Replace DLLP with LLP-GANLLP learning · CIFAR-100, bag sizes 16, 32, 64, and 128test error (%): 53.24(0.77) -> 50.95(0.67); 98.38(0.11) -> 56.44(0.78); 98.65(0.09) -> 64.37 (1.52); 98.98(0.08) -> 85.01 (1.81)Improved
Add entropy regularization to DLLP (ablation)DLLP on MNIST · MNISTinstance-level entropyNo clear effect

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