Research map / Large language models / Agents, code and math
Revolutionizing Training-Free NAS: Towards Efficient Automatic Proxy Discovery via Large Language Models
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Uses large language models to automatically discover efficient zero-cost proxies for training-free neural architecture search.
Problem
Training-free NAS relies on zero-cost proxies (ZCPs) computed at initialization to rank architectures, but existing hand-designed ZCPs are limited in effectiveness and generality.
Approach
Leverages large language models to automatically generate and refine zero-cost proxy code, discovering proxies that are more efficient and effective than manually designed ones.
Experimental setup
Baselines: hand-designed zero-cost proxies
Builds on: zero-cost proxies
Closest papers by meaning
- L-SWAG: Layer-Sample Wise Activation with Gradients Information for Zero-Shot NAS on Vision Transformers (CVPR 2025 · similarity 0.90)
- Auto-Prox: Training-Free Vision Transformer Architecture Search via Automatic Proxy Discovery (AAAI 2024 · similarity 0.89)
- RZ-NAS: Enhancing LLM-guided Neural Architecture Search via Reflective Zero-Cost Strategy (ICML 2025 · similarity 0.88)
- Training-free Neural Architecture Search through Variance of Knowledge of Deep Network Weights (CVPR 2025 · similarity 0.88)
- ParZC: Parametric Zero-Cost Proxies for Efficient NAS (AAAI 2025 · similarity 0.87)
- einspace: Searching for Neural Architectures from Fundamental Operations (NeurIPS 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.