LVO applies optimization-based feature visualization to latent diffusion models after disentangling their representations with sparse autoencoders, yielding recognizable concept images on a fine-tuned Stable Diffusion model that are clearer than those from entangled baselines.
Deconvolution and Checkerboard Artifacts
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 5roles
background 1polarities
background 1representative citing papers
Frozen features from vision foundation models enable a linear probe to outperform specialized AIGI detectors by over 30% on in-the-wild data due to emergent forgery knowledge from pre-training.
Flow matching achieves single-step pixel accuracy and 20-step perceptual quality for Sentinel-2 super-resolution, outperforming diffusion and Real-ESRGAN while enabling large-scale 2.5 m land-cover products.
GeomPrompt learns a task-driven geometric prompt from RGB alone to substitute for missing or degraded depth in frozen RGB-D semantic segmentation models, yielding up to +6.1 mIoU gains on SUN RGB-D while being faster than monocular depth estimators.
μ-FlowNet applies an attention U-Net to map flow fields in irregular microchannels, reporting dice score 0.9317 and IoU 0.8731 on test data while outperforming standard U-Net and T-Net.
citing papers explorer
-
Deep Dreams Are Made of This: Visualizing Monosemantic Features in Diffusion Models
LVO applies optimization-based feature visualization to latent diffusion models after disentangling their representations with sparse autoencoders, yielding recognizable concept images on a fine-tuned Stable Diffusion model that are clearer than those from entangled baselines.
-
Simplicity Prevails: The Emergence of Generalizable AIGI Detection in Visual Foundation Models
Frozen features from vision foundation models enable a linear probe to outperform specialized AIGI detectors by over 30% on in-the-wild data due to emergent forgery knowledge from pre-training.
-
Flow matching for Sentinel-2 super-resolution: implementation, application, and implications
Flow matching achieves single-step pixel accuracy and 20-step perceptual quality for Sentinel-2 super-resolution, outperforming diffusion and Real-ESRGAN while enabling large-scale 2.5 m land-cover products.
-
GeomPrompt: Geometric Prompt Learning for RGB-D Semantic Segmentation Under Missing and Degraded Depth
GeomPrompt learns a task-driven geometric prompt from RGB alone to substitute for missing or degraded depth in frozen RGB-D semantic segmentation models, yielding up to +6.1 mIoU gains on SUN RGB-D while being faster than monocular depth estimators.
-
$\mu$-FlowNet: A Deep Learning Approach for Mapping Flow Fields in Irregular Microchannels Using an Attention-based U-Net Encoder-Decoder Architecture
μ-FlowNet applies an attention U-Net to map flow fields in irregular microchannels, reporting dice score 0.9317 and IoU 0.8731 on test data while outperforming standard U-Net and T-Net.