AI Paper Atlas

Research map / Label-efficient and robust learning / Semi-supervised learning

A Unified Approach to Count-Based Weakly Supervised Learning

Vinay Shukla, Zhe Zeng, Kareem Ahmed, Guy Van den Broeck

NeurIPS 2023 · Label-efficient and robust learning › Semi-supervised learning · approach: Holistic semi-supervised pipelines

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 exact count loss instead of proportion loss PLLLP · Adult [0,1] bag size 8test AUC: 0.8884 -> 0.8985Improved
use count loss instead of LMMCMLLP · Magic [0,1] bag size 8test AUC: 0.9070 -> 0.9173Improved
use count loss instead of PLLLP · Magic [0,1] bag size 512test AUC: 0.8807 -> 0.9088Improved
use count loss at large bag sizeLLP · Adult [0,1] bag size 128test AUC: 0.8879 -> 0.8871No clear effect
use count loss instead of attention baselines at small training setMIL · MNIST N(10,2), 50 bagstest AUC: 0.775 -> 0.818Improved
use count loss with many training bagsMIL · MNIST N(50,10), 150 bagstest AUC: 0.970 -> 0.964Hurt
use count loss instead of attention baselinesMIL · Colon Canceraccuracy: 0.909 -> 0.915Improved
use count loss instead of Instance-MaxMIL instance-level · MNIST N(10,2), 50 bagsAUC: 0.4904 -> 0.8341Improved
use CL instead of CVIRPU learning · Binarized CIFARaccuracy: 82.3 -> 80.1Hurt
use CL-expect instead of CVIRPU learning · CIFAR Cat vs. Dogaccuracy: 73.3 -> 76.5Improved
use full distribution CL rather than expectation CL-expect (ablation)PU learning · CIFAR Cat vs. Dogaccuracy: 76.5 -> 74.8Hurt
use full distribution CL rather than CL-expect (ablation)PU learning · MNIST17accuracy: 98.7 -> 99.0Improved

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