Optimization

Straight to Zero: Why Linearly Decaying the Learning Rate to Zero Works Best for LLMs ICLR 2025 StreamGS: Online Generalizable Gaussian Splatting Reconstruction for Unposed Image Streams ICCV 2025 Structure-Conditional Minimum Bayes Risk Decoding EMNLP 2025 Structure Trumps Size: Rethinking Data Quality for LLM Reasoning EMNLP 2025 Taming LLMs with Gradient Grouping ACL 2025 Temporal Scaling Law for Large Language Models EMNLP 2025 The AdEMAMix Optimizer: Better, Faster, Older ICLR 2025 The Effect of SGD Batch Size on Autoencoder Learning: Sparsity, Sharpness, and Feature Learning JMLR 2025 Theoretical Insights into Fine-Tuning Attention Mechanism: Generalization and Optimization IJCAI 2025 The Surprising Effectiveness of Infinite-Width NTKs for Characterizing and Improving Model Training AAAI 2025 Tightening Robustness Verification of MaxPool-based Neural Networks via Minimizing the Over-Approximation Zone CVPR 2025 Tile-wise vs. Image-wise: Random-Tile Loss and Training Paradigm for Gaussian Splatting ICCV 2025 Tilted Sharpness-Aware Minimization ICML 2025 TITAN-Guide: Taming Inference-Time Alignment for Guided Text-to-Video Diffusion Models ICCV 2025 TKG-DM: Training-free Chroma Key Content Generation Diffusion Model CVPR 2025 TopV: Compatible Token Pruning with Inference Time Optimization for Fast and Low-Memory Multimodal Vision Language Model CVPR 2025 Toward Robust Neural Reconstruction from Sparse Point Sets CVPR 2025 Towards Adversarially Robust Dataset Distillation by Curvature Regularization AAAI 2025 Towards Better Robustness Against Natural Corruptions in Document Tampering Localization AAAI 2025 Towards Performance Consistency in Multi-Level Model Collaboration ICCV 2025 Towards Precise Scaling Laws for Video Diffusion Transformers CVPR 2025 Towards Stable and Storage-efficient Dataset Distillation: Matching Convexified Trajectory CVPR 2025 Training compute-optimal transformer encoder models EMNLP 2025 Training Dynamics Underlying Language Model Scaling Laws: Loss Deceleration and Zero-Sum Learning ACL 2025 Training with Fewer Bits: Unlocking Edge LLMs Training with Stochastic Rounding EMNLP 2025