HM-Bench is the first benchmark for MLLMs on hyperspectral images, showing models struggle with complex spatial-spectral reasoning and perform better with visual PCA images than textual reports.
Deep networks always grok and here is why
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
Persistent homology detects a sharp increase in maximum and total H1 persistence during grokking on modular arithmetic, offering a topological diagnostic that links representation geometry to generalization.
DISC extracts multi-statistic trajectories from diffusion denoising to both detect and classify types of distributional shifts in OOD data.
Self-supervised ReLU networks form substantially fewer linear regions than supervised models for comparable accuracy, with contrastive methods rapidly expanding regions and self-distillation consolidating them, enabling early geometric detection of representation collapse.
citing papers explorer
-
HM-Bench: A Comprehensive Benchmark for Multimodal Large Language Models in Hyperspectral Remote Sensing
HM-Bench is the first benchmark for MLLMs on hyperspectral images, showing models struggle with complex spatial-spectral reasoning and perform better with visual PCA images than textual reports.
-
Topological Signatures of Grokking
Persistent homology detects a sharp increase in maximum and total H1 persistence during grokking on modular arithmetic, offering a topological diagnostic that links representation geometry to generalization.
-
Beyond Binary Out-of-Distribution Detection: Characterizing Distributional Shifts with Multi-Statistic Diffusion Trajectories
DISC extracts multi-statistic trajectories from diffusion denoising to both detect and classify types of distributional shifts in OOD data.
-
Complexity of Linear Regions in Self-supervised Deep ReLU Networks
Self-supervised ReLU networks form substantially fewer linear regions than supervised models for comparable accuracy, with contrastive methods rapidly expanding regions and self-distillation consolidating them, enabling early geometric detection of representation collapse.