This paper proposes FaceMoE, an adaptation of Mixture of Experts (MoE) architecture for low-resolution face recognition, which addresses the challenges of poor feature extraction and domain gap.

The paper enhances the security of the PolyProtect biometric template protection method by proposing a key selection algorithm that significantly increases the difficulty of inverting protected face t…

This paper presents a memory-efficient training stack for Mixture-of-Experts (MoE) models, combining and specializing parallelism techniques for maximal efficiency.

A unified multi-task malware analysis framework based on Mixture of Experts (MoE) architectures is proposed for malware family classification, packed versus unpacked detection, and malware versus beni…

The paper develops a minimal dynamical model showing that adaptive softmax routing in Mixture-of-Experts (MoE) layers can undergo abrupt transitions to load imbalance via bifurcation mechanisms.

The paper analyzes the routing behavior of Mixtral MoE under benign and harmful prompts using activation and gradient signals, finding that safety-relevant routing is subtle, depth-dependent, and dist…

The paper proposes DAMEL, a dual-axis multi-expert learning algorithm that simultaneously reduces both prediction bias and variance in class-imbalanced learning by leveraging multiple experts across b…

This paper proposes a new router redesign for Mixture-of-Experts models using Manifold Power Iteration to align router rows with the principal singular directions of associated experts.

The paper introduces ProbMoE, a probabilistic routing framework that tackles the non-differentiability of top-$k$ routing in Mixture-of-Experts (MoE) models, achieving strong performance with improved…

This paper demonstrates that standard image classifiers can be interpreted as multiple-instance learning models, allowing for the recovery of spatial class evidence from image-level logits.

The paper proposes a privacy-preserving system for crowd monitoring that counts individuals across different locations and time periods using face recognition without ever revealing personal identitie…

VidPrism introduces a novel heterogeneous Mixture-of-Experts framework that specializes temporal processing by dividing labor among experts, achieving state-of-the-art performance in image-to-video tr…

The paper formalizes the problem of representation identifiability in supervised learning, showing that a representation property is identifiable if and only if it is constant across all possible fact…

This paper identifies scattering network architectures that maximize separation capacity for data with low intrinsic dimension by characterizing and bounding the separation capacity of general feature…

This paper extends gradient boosting to functions of vector inputs using a simple algorithm with histogram-based decision trees.

This paper examines the use of 2D barcodes on temporary travel documents to enable machine readability and automate biometric face verification while reducing storage capacity. It compares the perform…

This book provides a compact, derivation-oriented mathematical primer that connects major families of generative AI models, showing their underlying structural relationships.

The paper proposes DAG-MoE, a novel sparse Mixture-of-Experts framework that replaces standard weighted-sum aggregation with structural aggregation to enhance model performance and enable multi-step r…