SOTAVerified

Unsupervised Pre-training

Pre-training a neural network using unsupervised (self-supervised) auxiliary tasks on unlabeled data.

Papers

Showing 1120 of 265 papers

TitleStatusHype
SPACE: Your Genomic Profile Predictor is a Powerful DNA Foundation ModelCode1
PersonViT: Large-scale Self-supervised Vision Transformer for Person Re-IdentificationCode1
ConStyle v2: A Strong Prompter for All-in-One Image RestorationCode1
PEAC: Unsupervised Pre-training for Cross-Embodiment Reinforcement LearningCode1
BMRetriever: Tuning Large Language Models as Better Biomedical Text RetrieversCode1
Drop your Decoder: Pre-training with Bag-of-Word Prediction for Dense Passage RetrievalCode1
Unified Multi-modal Unsupervised Representation Learning for Skeleton-based Action UnderstandingCode1
METRA: Scalable Unsupervised RL with Metric-Aware AbstractionCode1
HIQL: Offline Goal-Conditioned RL with Latent States as ActionsCode1
Pre-training Contextualized World Models with In-the-wild Videos for Reinforcement LearningCode1
Show:102550
← PrevPage 2 of 27Next →

Benchmark Results

#ModelMetricClaimedVerifiedStatus
115RDLsAccuracy (%)95Unverified
29RDLsAccuracy (%)94Unverified
33 RMDLAccuracy (%)93Unverified
4CNNAccuracy (%)73Unverified
5RMDLAccuracy (%)0.1Unverified
#ModelMetricClaimedVerifiedStatus
1RMDL (30 RDLs)Sensitivity (VEB)90.69Unverified
2Sensitivity89.1Unverified
3RMDL 3 RDLsSensitivity0.87Unverified