SOTAVerified

Memorization

Papers

Showing 10011025 of 1088 papers

TitleStatusHype
The Information Complexity of Learning Tasks, their Structure and their Distance0
Understanding Feature Selection and Feature Memorization in Recurrent Neural Networks0
RepNet: Weakly Supervised Training of an Adversarial Reprojection Network for 3D Human Pose EstimationCode0
Learning to Infer Program SketchesCode0
Identity Crisis: Memorization and Generalization under Extreme Overparameterization0
Optimal Kronecker-Sum Approximation of Real Time Recurrent LearningCode0
Do We Train on Test Data? Purging CIFAR of Near-Duplicates0
Revisiting Sample Selection Approach to Positive-Unlabeled Learning: Turning Unlabeled Data into Positive rather than Negative0
State-Regularized Recurrent Neural Networks0
Enhancing human learning via spaced repetition optimizationCode1
How does Disagreement Help Generalization against Label Corruption?Code0
Detecting Overfitting of Deep Generative Networks via Latent RecoveryCode0
Learning to Remember More with Less MemorizationCode1
VideoMem: Constructing, Analyzing, Predicting Short-term and Long-term Video Memorability0
Limited Gradient Descent: Learning With Noisy Labels0
Leveraging Random Label Memorization for Unsupervised Pre-Training0
Auditing Data Provenance in Text-Generation ModelsCode0
Small ReLU networks are powerful memorizers: a tight analysis of memorization capacity0
Are You Sure YouWant To Do That? Classification with Interpretable Queries0
Memorization in Overparameterized Autoencoders0
Detecting Memorization in ReLU Networks0
Wide and Deep Learning for Peer-to-Peer Lending0
A Practical Approach to Sizing Neural Networks0
SIGUA: Forgetting May Make Learning with Noisy Labels More RobustCode0
Pumpout: A Meta Approach for Robustly Training Deep Neural Networks with Noisy Labels0
Show:102550
← PrevPage 41 of 44Next →

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1PaLM-540B (few-shot, k=5)Accuracy95.4Unverified
2Gopher-280B (few-shot, k=5)Accuracy80Unverified
3PaLM-62B (few-shot, k=5)Accuracy77.7Unverified