Beyond the Trial-and-Error Loop: Hybrid Projection and Automated Tuning for Distributed Training

cs.AI, cs.DC

Submitted: 2026-10-14

Updated: 2026-10-14

Comments: Accepted at NeurIPS 2026. 22 pages, 8 figures, 25 tables

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Sources

• DeepSeek-V3 Technical Report
• The Llama 3 Herd of Models
• Mixtral of Experts
• gpt-oss-120b & gpt-oss-20b Model Card
• Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism
• tritonBLAS: Triton-based Analytical Approach for GEMM Kernel Parameter Selection
• Qwen3 Technical Report
• Recursive Language Models

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• Agents' Last Exam
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems