Promptbreeder evolves both task prompts and the mutation prompts that improve them using LLMs, outperforming Chain-of-Thought and Plan-and-Solve on arithmetic and commonsense reasoning benchmarks.
hub Mixed citations
A Simple Framework for Contrastive Learning of Visual Representations
Mixed citation behavior. Most common role is background (57%).
63
Pith papers citing it
7,335
external citations · Pith
Background
57% of classified citations
abstract
This paper presents SimCLR: a simple framework for contrastive learning of visual representations. We simplify recently proposed contrastive self-supervised learning algorithms without requiring specialized architectures or a memory bank. In order to understand what enables the contrastive prediction tasks to learn useful representations, we systematically study the major components of our framework. We show that (1) composition of data augmentations plays a critical role in defining effective predictive tasks, (2) introducing a learnable nonlinear transformation between the representation and the contrastive loss substantially improves the quality of the learned representations, and (3) contrastive learning benefits from larger batch sizes and more training steps compared to supervised learning. By combining these findings, we are able to considerably outperform previous methods for self-supervised and semi-supervised learning on ImageNet. A linear classifier trained on self-supervised representations learned by SimCLR achieves 76.5% top-1 accuracy, which is a 7% relative improvement over previous state-of-the-art, matching the performance of a supervised ResNet-50. When fine-tuned on only 1% of the labels, we achieve 85.8% top-5 accuracy, outperforming AlexNet with 100X fewer labels.
hub tools
citation-role summary
background 4
method 3
citation-polarity summary
representative citing papers
Self-supervised ViTs show emergent semantic segmentation and 78.3% k-NN accuracy on ImageNet; DINO reaches 80.1% linear evaluation with ViT-Base.
LeVLJEPA is the first non-contrastive vision-language pretraining method that learns via cross-modal prediction without negatives, producing stronger dense features than contrastive baselines on VQA and segmentation tasks.
A self-supervised method pretrains an encoder on eight PSP images per view to learn generalizable subsurface scattering representations that transfer to relighting and dense footprint reconstruction on unseen complex objects.
UR-JEPA applies uniform rectifiability regularization via a smoothed Carleson square function to JEPA training, producing embeddings with 4-5 order PCA spectral drop at dimension 20-25 and lower seed variance than Gaussian regularization on Inet10, Galaxy10, and EuroSAT.
Predictive Entropy Maximization performs competitive blind source separation using only local error-driven and Hebbian updates derived from a surrogate entropy objective with spectral error bounds.
A two-stage contrastive teacher-student framework learns and then projects latent dynamics onto port-Hamiltonian submanifolds from partial observations.
SMA uses a submodular mutual information objective on data sets to deliver competitive zero-shot classification and retrieval performance on CLIP benchmarks with only tens of thousands of samples, orders of magnitude fewer than standard approaches.
NodePFN pre-trains on synthetic graphs with controllable homophily and causal feature-label models to achieve 71.27 average accuracy on 23 node classification benchmarks without graph-specific training.
SDTI lets models identify the correct target variable in datasets in a zero-shot setting using standard neural networks, beating baselines by 14% F1 on synthetic benchmarks.
BEiT pre-trains vision transformers via masked image modeling on visual tokens and reaches 83.2% ImageNet top-1 accuracy for the base model and 86.3% for the large model using only ImageNet-1K data.
DreamerV2 reaches human-level performance on 55 Atari games by learning behaviors inside a separately trained discrete-latent world model.
PGUDA uses pressure signals to train a teacher network that distills modality-invariant knowledge into an sEMG student via cross-modal distillation, reaching 58.08% cross-subject accuracy with only 5% labeled data for the teacher.
TactX learns a shared latent representation across three tactile sensor modalities via joint training on paired contacts, enabling zero-shot policy transfer and higher success on pick-and-place, insertion, wiping, and reorientation tasks.
vMFProto models each class as a mixture of von Mises-Fisher distributions on the sphere, giving part prototypes that are more consistent and less redundant than point prototypes.
On-policy self-distillation with sampled demonstrations reduces rollout diversity by amplifying existing probability gaps in the base model, unlike ideal RL which preserves ratios among correct outputs.
SimPhysNet achieves 96.06% accuracy classifying laser welding penetration states using self-supervised contrastive learning with a physics-informed neural network and prototypical networks on only 200 labeled images.
Introduces Hell-Char, PaLit-Char and Med-Char datasets plus a similarity-weighted supervised contrastive loss and lacuna augmentation to learn diachronic embeddings for ancient Greek letterforms.
A neural marking scheme trained with contrastive learning tightens constraints on σ8 by 2.9× and Ωm by 1.8× over classical marks at k_max=0.2 h/Mpc while breaking their degeneracy at the Fisher level.
DALE-CT, a 2D LeJEPA model with depth-aware dual supervision, reaches 0.833 Macro AUROC on multi-abnormality detection in CT and approaches 3D SOTA performance using less data and no textual supervision.
FBCC introduces unsupervised continual clustering via a teacher-student forward-backward distillation process that outperforms baselines on clustering accuracy while reducing catastrophic forgetting.
PLCI infers posterior distributions over latent cross-view counterparts using instance reliability and prototype transport to jointly address incomplete views and noisy correspondences in multi-view clustering.
Sampling pairs directly with auxiliary information for higher inclusion probabilities on informative pairs yields near-full pairwise loss performance at reduced computational cost.
DexHoldem is a new benchmark providing 1,470 teleoperated demonstrations across 14 manipulation primitives, plus standardized tests for dexterous policy execution and agentic perception in a physical Texas Hold'em setting.
citing papers explorer
-
Promptbreeder: Self-Referential Self-Improvement Via Prompt Evolution
Promptbreeder evolves both task prompts and the mutation prompts that improve them using LLMs, outperforming Chain-of-Thought and Plan-and-Solve on arithmetic and commonsense reasoning benchmarks.
-
Emerging Properties in Self-Supervised Vision Transformers
Self-supervised ViTs show emergent semantic segmentation and 78.3% k-NN accuracy on ImageNet; DINO reaches 80.1% linear evaluation with ViT-Base.
-
LeVLJEPA: End-to-End Vision-Language Pretraining Without Negatives
LeVLJEPA is the first non-contrastive vision-language pretraining method that learns via cross-modal prediction without negatives, producing stronger dense features than contrastive baselines on VQA and segmentation tasks.
-
From Phase to Phenomenon: Self-Supervised Learning of Subsurface Scattering with Minimal Phase-shift Inputs
A self-supervised method pretrains an encoder on eight PSP images per view to learn generalizable subsurface scattering representations that transfer to relighting and dense footprint reconstruction on unseen complex objects.
-
UR-JEPA: Uniform Rectifiability as a Regularizer for Joint-Embedding Predictive Architectures
UR-JEPA applies uniform rectifiability regularization via a smoothed Carleson square function to JEPA training, producing embeddings with 4-5 order PCA spectral drop at dimension 20-25 and lower seed variance than Gaussian regularization on Inet10, Galaxy10, and EuroSAT.
-
Normative Networks for Source Separation via Local Plasticity and Dendritic Computation
Predictive Entropy Maximization performs competitive blind source separation using only local error-driven and Hebbian updates derived from a surrogate entropy objective with spectral error bounds.
-
Identify Then Project: Contrastive Learning of Latent Dynamics from Partial Observations with Port-Hamiltonian Structure
A two-stage contrastive teacher-student framework learns and then projects latent dynamics onto port-Hamiltonian submanifolds from partial observations.
-
SMA: Submodular Modality Aligner For Data Efficient Multimodal Learning
SMA uses a submodular mutual information objective on data sets to deliver competitive zero-shot classification and retrieval performance on CLIP benchmarks with only tens of thousands of samples, orders of magnitude fewer than standard approaches.
-
Learning Posterior Predictive Distributions for Node Classification from Synthetic Graph Priors
NodePFN pre-trains on synthetic graphs with controllable homophily and causal feature-label models to achieve 71.27 average accuracy on 23 node classification benchmarks without graph-specific training.
-
Self-Directed Task Identification
SDTI lets models identify the correct target variable in datasets in a zero-shot setting using standard neural networks, beating baselines by 14% F1 on synthetic benchmarks.
-
BEiT: BERT Pre-Training of Image Transformers
BEiT pre-trains vision transformers via masked image modeling on visual tokens and reaches 83.2% ImageNet top-1 accuracy for the base model and 86.3% for the large model using only ImageNet-1K data.
-
Mastering Atari with Discrete World Models
DreamerV2 reaches human-level performance on 55 Atari games by learning behaviors inside a separately trained discrete-latent world model.
-
PGUDA: Pressure-Guided Unsupervised Domain Adaptation with Cross-Modal Knowledge Distillation for sEMG-Based Gesture Recognition
PGUDA uses pressure signals to train a teacher network that distills modality-invariant knowledge into an sEMG student via cross-modal distillation, reaching 58.08% cross-subject accuracy with only 5% labeled data for the teacher.
-
TactX: Learning Shared Tactile Representations Across Diverse Sensors
TactX learns a shared latent representation across three tactile sensor modalities via joint training on paired contacts, enabling zero-shot policy transfer and higher success on pick-and-place, insertion, wiping, and reorientation tasks.
-
Beyond Points: Spherical Distributional Part Prototypes for Interpretable Classification
vMFProto models each class as a mixture of von Mises-Fisher distributions on the sphere, giving part prototypes that are more consistent and less redundant than point prototypes.
-
On-Policy Self-Distillation with Sampled Demonstrations Reduces Output Diversity
On-policy self-distillation with sampled demonstrations reduces rollout diversity by amplifying existing probability gaps in the base model, unlike ideal RL which preserves ratios among correct outputs.
-
A welding penetration prediction model for laser welding process based on self-supervised learning using physics-informed neural networks
SimPhysNet achieves 96.06% accuracy classifying laser welding penetration states using self-supervised contrastive learning with a physics-informed neural network and prototypical networks on only 200 labeled images.
-
Learning Diachronic Representations of Ancient Greek Letterforms
Introduces Hell-Char, PaLit-Char and Med-Char datasets plus a similarity-weighted supervised contrastive loss and lacuna augmentation to learn diachronic embeddings for ancient Greek letterforms.
-
Interpretable Neural Marked Statistics for Cosmological Inference
A neural marking scheme trained with contrastive learning tightens constraints on σ8 by 2.9× and Ωm by 1.8× over classical marks at k_max=0.2 h/Mpc while breaking their degeneracy at the Fisher level.
-
DALE-CT: Depth-Aware Foundation Models for Computed Tomography
DALE-CT, a 2D LeJEPA model with depth-aware dual supervision, reaches 0.833 Macro AUROC on multi-abnormality detection in CT and approaches 3D SOTA performance using less data and no textual supervision.
-
Unsupervised Continual Clustering via Forward-Backward Knowledge Distillation
FBCC introduces unsupervised continual clustering via a teacher-student forward-backward distillation process that outperforms baselines on clustering accuracy while reducing catastrophic forgetting.
-
Robust Multi-view Clustering against Imperfect Information
PLCI infers posterior distributions over latent cross-view counterparts using instance reliability and prototype transport to jointly address incomplete views and noisy correspondences in multi-view clustering.
-
Doing well with less! On Sampling Techniques for Empirical Pairwise Loss Estimation/Minimization
Sampling pairs directly with auxiliary information for higher inclusion probabilities on informative pairs yields near-full pairwise loss performance at reduced computational cost.
-
DexHoldem: Playing Texas Hold'em with Dexterous Embodied System
DexHoldem is a new benchmark providing 1,470 teleoperated demonstrations across 14 manipulation primitives, plus standardized tests for dexterous policy execution and agentic perception in a physical Texas Hold'em setting.
-
Self-organized MT Direction Maps Emerge from Spatiotemporal Contrastive Optimization
Direction maps and pinwheel structures in MT emerge spontaneously when a spatiotemporal deep network is trained on videos with contrastive self-supervised learning and spatial regularization.
-
CalibFree: Self-Supervised View Feature Separation for Calibration-Free Multi-Camera Multi-Object Tracking
CalibFree enables calibration-free multi-camera tracking via self-supervised feature separation through single-view distillation and cross-view reconstruction, reporting 3% higher accuracy and 7.5% better F1 on tested datasets.
-
An Interpretable and Scalable Framework for Evaluating Large Language Models
A majorization-minimization framework turns IRT into scalable matrix factorization subproblems for LLM evaluation, delivering orders-of-magnitude speedups with identifiability guarantees.
-
Velox: Learning Representations of 4D Geometry and Appearance
Velox compresses dynamic point clouds into latent tokens that support geometry via 4D surface modeling and appearance via 3D Gaussians, showing strong results on video-to-4D generation, tracking, and image-to-4D cloth simulation.
-
ShapeY: A Principled Framework for Measuring Shape Recognition Capacity via Nearest-Neighbor Matching
ShapeY is a benchmark dataset and nearest-neighbor protocol that measures shape-based recognition in vision models, revealing that even state-of-the-art networks fail to generalize consistently across 3D viewpoints and non-shape appearance changes.
-
StarCLR: Contrastive Learning Representation for Astronomical Light Curves
StarCLR pretrains on TESS light curves via contrastive learning on overlapping subsequences and improves variable star classification F1 scores over scratch-trained models when fine-tuned on TESS, ZTF, and Gaia.
-
Only Brains Align with Brains: Cross-Region Alignment Patterns Expose Limits of Normative Models
Alignment pattern analysis reveals that models aligned to individual brain ROIs do not reproduce the stable cross-region alignment profiles observed across human subjects.
-
Self-supervised Pretraining of Cell Segmentation Models
DINOCell achieves a SEG score of 0.784 on LIVECell by self-supervised domain adaptation of DINOv2, improving 10.42% over SAM-based models and showing strong zero-shot transfer.
-
Zero-shot World Models Are Developmentally Efficient Learners
A zero-shot visual world model trained on one child's experience achieves broad competence on physical understanding benchmarks while matching developmental behavioral patterns.
-
Masked Contrastive Pre-Training Improves Music Audio Key Detection
Masked contrastive pre-training on music audio yields representations that achieve SOTA key detection performance in the supervised setting without sophisticated augmentations.
-
Unsupervised domain adaptation for radioisotope identification in gamma spectroscopy
Unsupervised domain adaptation via feature alignment raises radioisotope identification accuracy on real LaBr3 gamma spectra from 0.754 to 0.904 for models trained only on synthetic data.
-
SSL4RL: Revisiting Self-supervised Learning as Intrinsic Reward for Visual-Language Reasoning
SSL4RL reformulates self-supervised learning objectives into dense, verifiable reward signals for RL-based fine-tuning of vision-language models, yielding performance gains on reasoning benchmarks.
-
Painless Activation Steering: An Automated, Lightweight Approach for Post-Training Large Language Models
PAS automates activation steering for LLMs using labeled data to improve behavior control on tasks like bias and alignment, with gains over ICL and SFT but limited effect on intelligence tasks.
-
Revisiting Feature Prediction for Learning Visual Representations from Video
V-JEPA models trained only on feature prediction from 2 million public videos achieve 81.9% on Kinetics-400, 72.2% on Something-Something-v2, and 77.9% on ImageNet-1K using frozen ViT-H/16 backbones.
-
Demystifying CLIP Data
MetaCLIP curates balanced 400M-pair subsets from CommonCrawl that outperform CLIP data, reaching 70.8% zero-shot ImageNet accuracy on ViT-B versus CLIP's 68.3%.
-
Template-assisted Contrastive Learning of Task-oriented Dialogue Sentence Embeddings
TaDSE learns dialogue sentence embeddings via template-guided self-supervised contrastive learning plus synthetic slot-filling augmentation and reports gains on five downstream benchmarks.
-
STST-JEPA: Shallow-Target Spatio-Temporal Joint Embedding Prediction Architecture For EEG Self-Supervised Learning
A JEPA-style EEG foundation model with shallow EMA targets plus light reconstruction reaches strong multi-task transfer and 3.06-year validation age MAE on a large multi-site corpus.
-
LeNEPA: No-Augmentation Next-Latent Prediction for Time-Series Representation Learning
LeNEPA proposes a no-augmentation next-latent prediction recipe that maintains frozen-probe performance across ECG and synthetic diagnostic time-series datasets under fixed-recipe conditions where a tuned JEPA baseline degrades.
-
Data Selection Through Iterative Self-Filtering for Vision-Language Settings
An iterative bootstrapped self-filtering approach selects balanced clean and diverse subsets from noisy vision-language datasets to train improved CLIP models.
-
CoughPhase-CLR: Designing an acoustics-informed foundation model for coughing sound classification
CoughPhase-CLR uses cough physiological phases to build contrastive positive pairs, outperforming random cropping on downstream tasks including COVID-19 detection and COPD classification.
-
CoRe-MoE: Contrastive Reweighted Mixture of Experts for Multi-Terrain Humanoid Locomotion with Gait Adaptation
CoRe-MoE uses a two-stage RL framework with contrastive reweighting in a Mixture-of-Experts architecture to enable gait transitions and multi-terrain adaptation for humanoid locomotion.
-
Enhancing the Socioeconomic Understanding of Foundation Models with Urban Mobility
MobFusion fuses mobility networks into foundation models via three designs and reports improved performance on income, density, and crime prediction tasks using data from three U.S. metropolitan areas.
-
Representation Without Reward: A JEPA Audit for LLM Fine-Tuning
An empirical audit of 22 JEPA-style training auxiliaries on Llama-3.2-1B fine-tuning for regex generation finds no statistically significant task improvement after multiple-testing correction, even when auxiliaries visibly alter hidden-state geometry.
-
Dynamical Predictive Modelling of Cardiovascular Disease Progression Post-Myocardial Infarction via ECG-Trained Artificial Intelligence Model
A contrastive-learning ECG foundation model with multitask heads predicts post-MI outcomes better than training from scratch (AUC 0.794 vs 0.608).
-
Information theoretic underpinning of self-supervised learning by clustering
SSL clustering is derived as KL-divergence optimization where a teacher-distribution constraint normalizes via inverse cluster priors and simplifies to batch centering by Jensen's inequality.
-
Handling and Interpreting Missing Modalities in Patient Clinical Trajectories via Autoregressive Sequence Modeling
Autoregressive transformer modeling with missingness-aware contrastive pre-training outperforms baselines on MIMIC-IV and eICU benchmarks and mitigates divergent behavior from removed modalities in clinical trajectories.