Modular Norm RandOpt: Population-Efficient Ensembling through Architecture-Aware Perturbations
cs.LG
Submitted: 2026-09-22
Updated: 2026-09-23
Comments: Preprint. Project page: https://kiratoyoshihara.github.io/Modular-Norm-RandOpt-page/
Project page: https://kiratoyoshihara.github.io/Modular-Norm-RandOpt-page
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Program Synthesis with Large Language Models
- Layer Normalization
- Constitutional AI: Harmlessness from AI Feedback
- Training Verifiers to Solve Math Word Problems
- Gemma 3 Technical Report
- Snapshot Ensembles: Train 1, get M for free
- Mistral 7B
- Not the Dimension, the Norm: What Matters in Gradient-Free Weight Perturbation of Language Models
- The Llama 3 Herd of Models
- Qwen2.5 Technical Report
- Evolution Strategies as a Scalable Alternative to Reinforcement Learning
- GLU Variants Improve Transformer
- Olmo 3
- LLaMA: Open and Efficient Foundation Language Models
- Llama 2: Open Foundation and Fine-Tuned Chat Models
- Self-Consistency Improves Chain of Thought Reasoning in Language Models
- Different Layers, Different Manifolds: Module-Wise Weight-Space Geometry in Transformer Optimization
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks