Sep 2026· International Journal of Data Science and Analysis· Vol 22· 0 citations· 47 references
TL;DR
MGOML-NGF, a geometry-aware oversampling technique that integrates neural geometry fields with a meta-learning control mechanism to enable adaptive minority-sample generation, indicates that neural geometry-aware oversampling can improve minority-class representation while reducing artifacts caused by linear interpolation in high-dimensional feature spaces.
These results identify conditions under which the resulting predictor exploits local geometry and attains the aggregate minimax rate, and derive an in-context generalization bound for near empirical risk minimizers over this class.
MoLiegroup is introduced, a framework that embeds multiple Lie-group priors as specialized kernel experts and adaptively fuses them via a geometry-aware gating mechanism to learn equivariance-inspired geometric features and supplies a practical, architecture-agnostic geometric inductive bias that improves expressivity...
Yao-Xian Yang, Guipeng Lan, Shuai Xiao et al.· IEEE Transactions on Neural...· 0 citations
This work introduces a novel framework, Gaussian Bridge Consistency (GBC), to address challenges of semi-supervised learning by constructing semantic interpolation paths between unlabeled samples and high-quality class anchors, and proposes BridgeMix, a confidence-aware feature mixing strategy that interpolates both sa...
Hong-Yang He, Xin-Yuan Song, Yan Zhong et al.· 1 citation
GeoTTER is a novel framework that redefines optimal transport in the realm of zero-shot classification by incorporating two key techniques to alleviate high-frequency label jaggedness and correct coherent angular drift, achieving a multi-objective optimization that respects both global distribution constraints and late...
W. Lee, Rudrasis Chakraborty, V. Lokhande· 0 citations
In molecular property prediction, unsupervised graph contrastive learning (GCL) is pivotal when obtaining labeled data is costly. However, conventional augmentation-free approaches based on parameter perturbation mainly rely on full-rank isotropic noise, which can introduce uncontrolled perturbations and impair semanti...
Xue-Yi Tan· Applied and Computational En...· 0 citations
DMT-Dens is a parametric manifold-visualization method built on a latent-token Transformer encoder that optimizes a loss based on the Pearson correlation between k-nearest-neighbor log-radius estimates in the processed input and two-dimensional embedding spaces to preserve density.
Ruizhe Wang, Yi-Xuan Dong, Bolin Yang et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
