Cover Image

Geometric Metrics for MoE Specialization: From Fisher Information to Early Failure Detection

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Guo, Dongxin, Wu, Jikun, Yiu, Siu Ming
Format:	Preprint
Published:	2026
Subjects:	Artificial Intelligence 62B10, 68T07 I.2.6; G.3
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

SigGate-GT: Taming Over-Smoothing in Graph Transformers via Sigmoid-Gated Attention
by: Guo, Dongxin, et al.
Published: (2026)

When Do Early-Exit Networks Generalize? A PAC-Bayesian Theory of Adaptive Depth
by: Guo, Dongxin, et al.
Published: (2026)

SafeAnchor: Preventing Cumulative Safety Erosion in Continual Domain Adaptation of Large Language Models
by: Guo, Dongxin, et al.
Published: (2026)

Closing the Theory-Practice Gap in Spiking Transformers via Effective Dimension
by: Guo, Dongxin, et al.
Published: (2026)

Parameter-Efficient Neuroevolution for Diverse LLM Generation: Quality-Diversity Optimization via Prompt Embedding Evolution
by: Guo, Dongxin, et al.
Published: (2026)

Jacobian-Velocity Bounds for Deployment Risk Under Covariate Drift
by: Landers, Jonathan R.
Published: (2026)

Constant-Target Energy Matching: A Unified Framework for Continuous and Discrete Density Estimation
by: Zeng, Zhijun, et al.
Published: (2026)

Robust Probabilistic Load Forecasting for a Single Household: A Comparative Study from SARIMA to Transformers on the REFIT Dataset
by: Manoj, Midhun
Published: (2025)

Evaluating the Quality of the Quantified Uncertainty for (Re)Calibration of Data-Driven Regression Models
by: Wibbeke, Jelke, et al.
Published: (2025)

Machine Collaboration
by: Liu, Qingfeng, et al.
Published: (2021)

EEG-EyeTrack: A Benchmark for Time Series and Functional Data Analysis with Open Challenges and Baselines
by: Afonso, Tiago Vasconcelos, et al.
Published: (2025)

Latent-Autoregressive GP-VAE Language Model
by: Ruffenach, Yves
Published: (2025)

Seeking SOTA: Time-Series Forecasting Must Adopt Taxonomy-Specific Evaluation to Dispel Illusory Gains
by: Saqur, Raeid, et al.
Published: (2026)

EvoPref: Multi-Objective Evolutionary Optimization Discovers Diverse LLM Alignments Beyond Gradient Descent
by: Guo, Dongxin, et al.
Published: (2026)

The Matching Principle: A Geometric Theory of Loss Functions for Nuisance-Robust Representation Learning
by: Rajput, Vishal
Published: (2026)

Multiple data-driven missing imputation
by: Kavun, Sergii
Published: (2025)

PatchFormer: A Patch-Based Time Series Foundation Model with Hierarchical Masked Reconstruction and Cross-Domain Transfer Learning for Zero-Shot Multi-Horizon Forecasting
by: Imanov, Olaf Yunus Laitinen, et al.
Published: (2026)

GT-PCA: Effective and Interpretable Dimensionality Reduction with General Transform-Invariant Principal Component Analysis
by: Heinrichs, Florian
Published: (2024)

Revisiting Unbiased Implicit Variational Inference
by: Pielok, Tobias, et al.
Published: (2025)

Gradient Descent as Implicit EM in Distance-Based Neural Models
by: Oursland, Alan
Published: (2025)

CDFlow: Building Invertible Layers with Circulant and Diagonal Matrices
by: Feng, Xuchen, et al.
Published: (2025)

Temporal Functional Circuits: From Spline Plots to Faithful Explanations in KAN Forecasting
by: Mysore, Naveen
Published: (2026)

Forecasting Coccidioidomycosis (Valley Fever) in Arizona: A Graph Neural Network Approach
by: Sarabi, Ali, et al.
Published: (2025)

Adam Improves Muon: Adaptive Moment Estimation with Orthogonalized Momentum
by: Zhang, Minxin, et al.
Published: (2026)

metabeta -- A fast neural model for Bayesian mixed-effects regression
by: Kipnis, Alex, et al.
Published: (2025)

Autoencoders in Function Space
by: Bunker, Justin, et al.
Published: (2024)

Data-Driven Discovery of Feature Groups in Clinical Time Series
by: Sergeev, Fedor, et al.
Published: (2025)

Randomized Spline Trees for Functional Data Classification: Theory and Application to Environmental Time Series
by: Riccio, Donato, et al.
Published: (2024)

Semi-Implicit Variational Inference via Kernelized Path Gradient Descent
by: Pielok, Tobias, et al.
Published: (2025)

From Tail Universality to Bernstein-von Mises: A Unified Statistical Theory of Semi-Implicit Variational Inference
by: Plummer, Sean
Published: (2025)

Deriving Decoder-Free Sparse Autoencoders from First Principles
by: Oursland, Alan
Published: (2026)

Multi-Hypothesis Prediction for Portfolio Optimization: A Structured Ensemble Learning Approach to Risk Diversification
by: Dominguez, Alejandro Rodriguez, et al.
Published: (2025)

From 'What-is' to 'What-if' in Human-Factor Analysis: A Post-Occupancy Evaluation Case
by: Chen, Xia, et al.
Published: (2025)

Ultrahigh-Q chiral resonances empowered by multi-head attention deep learning
by: Zhang, Cong, et al.
Published: (2025)

MRMS-Net and LMRMS-Net: Scalable Multi-Representation Multi-Scale Networks for Time Series Classification
by: Alagöz, Celal, et al.
Published: (2026)

ProactBench: Beyond What The User Asked For
by: Harfi, Sepehr, et al.
Published: (2026)

When Can Human-AI Teams Outperform Individuals? Tight Bounds with Impossibility Guarantees
by: Guo, Dongxin, et al.
Published: (2026)

GRALIS: A Unified Canonical Framework for Linear Attribution Methods via Riesz Representation
by: Fanale, Raimondo
Published: (2026)

Less is More: Strategic Expert Selection Outperforms Ensemble Complexity in Traffic Forecasting
by: Guettala, Walid, et al.
Published: (2025)

Provable Adversarial Robustness in In-Context Learning
by: Zhang, Di
Published: (2026)