Ghosted Layers: Unconstrained Activation Alignment for Recovering Layer-Pruned LLMs

Robust Multi-Agent LLMs under Byzantine Faults

Revisiting 16-bit Neural Network Training: A Practical Approach for Resource-Limited Learning

Locality-Aware Redundancy Pruning for LLM Depth Compression

Rethinking Layer Redundancy: Calibration Matters More Than Search in LLM Depth Pruning

Ghosted Layers: Unconstrained Activation Alignment for Recovering Layer-Pruned LLMs

Spectrum-Shaping Regularization for Neural Network Training Robust to Low-Rank Pruning

Weight Concentration Regularization for Improving Pruning Robustness Under High Sparsity

Revisiting 16-bit Neural Network Training: A Practical Approach for Resource-Limited Learning

Robust Neural Pruning with Gradient Sampling Optimization for Residual Neural Networks

Robust Multi-Agent LLMs under Byzantine Faults

Sharpness-Aware Minimization with Z-Score Gradient Filtering

On How Muon Reshapes Skill Learning Dynamics

Asynchronous Sharpness-Aware Minimization for Fast and Accurate Deep Learning

Why Does Stochastic Gradient Descent Slow Down in Low-Precision Training?

Hyperparameter-Free Auto-Scaled Gradient Normalization via Global Standard Deviation Dynamics

Fast Fourier Transform-Based Spectral and Temporal Gradient Filtering for Differential Privacy

Stochastic Gradient Sampling for Enhancing Neural Networks Training

ZNorm: Z-Score Gradient Normalization Accelerating Skip-Connected Network Training Without Architectural Modification

Mitigating Gradient Overlap in Deep Residual Networks with Gradient Normalization for Improved Non-Convex Optimization

MedCLM: Learning to Localize and Reason via a CoT-Curriculum in Medical Vision-Language Models

Merging Multiple Independently Trained Neural Networks Based on Genetic Algorithm

Extreme Solar Flare Prediction Using Residual Networks with HMI Magnetograms and Intensitygrams

Analysis and Predictive Modeling of Solar Coronal Holes Using Computer Vision and ARIMA-LSTM Networks

Uncertainty Estimation for Tumor Prediction with Unlabeled Data