From the 1 of 4 linked papers with an AI index.
4 papers
From Preimage Search To Source-Grounded Feature Inversion
Kaixiang Shu
The paper introduces a source‑grounded feature inversion method that reconstructs inputs by conditioning on the local network geometry of the original sample, using closed‑form mat…
Adjoint Inversion Reveals Holographic Superposition and Destructive Interference in CNN Classifiers
Kaixiang Shu
A foundational assumption in CNN interpretability -- that deep encoders suppress background pixels while classifiers merely select from a cleaned feature pool (the Spatial Funnel H…
CryoLVM: Self-supervised Learning from Cryo-EM Density Maps with Large Vision Models
Weining Fu, Kai Shu, Kui Xu +1
Cryo-electron microscopy (cryo-EM) has revolutionized structural biology by enabling near-atomic-level visualization of biomolecular assemblies. However, the exponential growth in…
Spatial Information Bottleneck for Interpretable Visual Recognition
Kaixiang Shu, Kai Meng, Junqin Luo
Deep neural networks typically learn spatially entangled representations that conflate discriminative foreground features with spurious background correlations, thereby undermining…