Free Reads
Sign in to view your remaining parses.
Tag Filter
Contrastive Learning Framework
A Simple Framework for Contrastive Learning of Visual Representations
Published:2/14/2020
Contrastive Learning FrameworkVisual Representation LearningSelf-Supervised Learning AlgorithmsImage Classification Benchmark DatasetData Augmentation Composition Strategy
The paper introduces SimCLR, a simplified framework for contrastive learning that eliminates the need for complex architectures or memory banks. Key findings highlight the importance of data augmentation combinations, a learnable nonlinear transformation, and larger batch sizes,
03
PvP: Data-Efficient Humanoid Robot Learning with Proprioceptive-Privileged Contrastive Representations
Published:12/15/2025
Humanoid Whole-Body ControlHumanoid Robot LearningContrastive Learning FrameworkState Representation Learning MethodsData-Efficient Reinforcement Learning
The paper introduces the PvP framework, addressing sample inefficiency in humanoid robot control by leveraging the complementarity of proprioceptive and privileged states. It improves learning efficiency without manual data augmentation, significantly enhancing performance in vel
04
DiffCL: A Diffusion-Based Contrastive Learning Framework with Semantic
Alignment for Multimodal Recommendations
Published:1/2/2025
Multimodal Recommendation SystemsDiffusion ModelsContrastive Learning FrameworkSemantic AlignmentGraph-Based Feature Enhancement
DiffCL is a diffusionbased contrastive learning framework for multimodal recommendation that reduces noise via diffusiongenerated views, aligns crossmodal semantics with stable ID embeddings, and alleviates data sparsity using an itemitem graph, enhancing recommendation accur
02
Towards Low-Latency and Adaptive Ransomware Detection Using Contrastive
Learning
Published:10/25/2025
Ransomware DetectionContrastive Learning FrameworkNeural Architecture SearchHardware Performance Counters AnalysisLow-Latency Malware Detection
This work integrates contrastive learning with neural architecture search and hardware performance data for lowlatency, adaptive ransomware detection, improving accuracy by 16.1% and response speed by six times while maintaining robustness against evasive attacks.
04