Log på
indsigt - Representation Learning
Poly-View Contrastive Learning: Maximizing Information for Improved Representation Learning
Poly-view contrastive learning maximizes information by increasing view multiplicity, leading to improved representation learning.
MIM4D: Masked Modeling for Autonomous Driving Representation Learning
MIM4D proposes a novel pre-training paradigm based on dual masked image modeling (MIM) for autonomous driving representation learning, achieving state-of-the-art performance on the nuScenes dataset.
Self-supervised Representation Learning From Random Data Projectors
Learning high-quality data representations from randomness is a feasible and plausible alternative when transformation invariance is challenging.
DecisionNCE: Embodied Multimodal Representations via Implicit Preference Learning
DecisionNCE는 결정을 위한 다중 모달 표현을 내재적 선호 학습을 통해 효과적으로 추출하는 통합된 표현 학습 프레임워크를 제안합니다.
K-Link: Enhancing Representation Learning in Multivariate Time-Series Data with Knowledge-Link Graph from LLMs
Knowledge-Link Graph from LLMs improves representation learning in MTS data.
Produkter | Ressourcer
© 2024 by Linnk AI