A variational latent bottleneck with KL regularization and a dynamic binary mask based on saliency produces model-specific features that keep high accuracy for one classifier but drop others below 2% on CIFAR-100 with over 45x suppression.
Title resolution pending
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
roles
background 2polarities
background 2representative citing papers
A weighted similarity ensemble unifies user-item and item-item collaborative filtering using shared embeddings to deliver competitive top-N recommendations without extra fine-tuning.
WKB analysis of the Teukolsky equation establishes a quasinormal-mode to greybody-factor correspondence for Kerr black holes that holds in the eikonal limit for gravitational perturbations and matches numerics at high angular momentum.
Longitudinal evaluation over yearly Android app slices shows temporal drift reduces adversarial robustness of malware detectors, with expanding-window retraining providing partial mitigation but not full recovery.
An integrated IoT and CNN system detects cracks in additive manufacturing with 99.54% accuracy and supports predictive maintenance via digital twins.
citing papers explorer
-
Variational Feature Compression for Model-Specific Representations
A variational latent bottleneck with KL regularization and a dynamic binary mask based on saliency produces model-specific features that keep high accuracy for one classifier but drop others below 2% on CIFAR-100 with over 45x suppression.
-
Collaborative Filtering Through Weighted Similarities of User and Item Embeddings
A weighted similarity ensemble unifies user-item and item-item collaborative filtering using shared embeddings to deliver competitive top-N recommendations without extra fine-tuning.
-
Quasinormal mode/grey-body factor correspondence for Kerr black holes
WKB analysis of the Teukolsky equation establishes a quasinormal-mode to greybody-factor correspondence for Kerr black holes that holds in the eikonal limit for gravitational perturbations and matches numerics at high angular momentum.
-
Adversarial Vulnerability Under Temporal Concept Drift: A Longitudinal Study of Android Malware Detection
Longitudinal evaluation over yearly Android app slices shows temporal drift reduces adversarial robustness of malware detectors, with expanding-window retraining providing partial mitigation but not full recovery.
-
IoT-Enhanced CNN-Based Labelled Crack Detection for Additive Manufacturing Image Annotation in Industry 4.0
An integrated IoT and CNN system detects cracks in additive manufacturing with 99.54% accuracy and supports predictive maintenance via digital twins.