Research map / Architectures and efficiency / RNNs and normalization
Scaling Law for Large Wireless Models
Ziheng Liu, Jiayi Zhang, Haoyu Wang, Bokai Xu, Chen Zhang, Yiyang Zhu, Enyu Shi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add channel heterogeneity and discretization granularity to classical scaling law | LWM scaling law · DeepMIMO unseen config | prediction accuracy: +32.31% prediction accuracy | Improved |
| add granularity beta on top of heterogeneity-only law (ablation) | LWM scaling law · DeepMIMO unseen config | prediction accuracy: +33.83% vs alpha-only | Improved |
| add heterogeneity alpha on top of granularity-only law (ablation) | LWM scaling law · DeepMIMO unseen config | prediction accuracy: +23.38% vs beta-only | Improved |
| refine granularity (channel-aware) at fixed data | LWM scaling law · DeepMIMO | training data needed: over 90% less training data for given error | Improved |
| increase LoS fraction alpha | LWM scaling law · DeepMIMO | irreducible test loss | Improved |
| use LWM representations instead of raw signal | LWM scaling law · downstream tasks (Rural/Urban) | F1: e.g. L/N 0.348 -> 0.955 (Rural-1) | Improved |
| use wireless law instead of classical law under +/-20% parameter perturbation | LWM scaling law · DeepMIMO | prediction error: <5.3% vs >14.2% error | Improved |
Closest papers by meaning
- Physics-Informed Generative Modeling of Wireless Channels (ICML 2025 · similarity 0.83)
- Power Lines: Scaling laws for weight decay and batch size in LLM pre-training (NeurIPS 2025 · similarity 0.82)
- Predicting Emergent Abilities with Infinite Resolution Evaluation (ICLR 2024 · similarity 0.82)
- TSNBench: Benchmarking LLM Proficiency in Time-Sensitive Networking (NeurIPS 2026 · similarity 0.82)
- A Communication-Theoretic Framework for LLM Agents: Cost-Aware Adaptive Reliability (NeurIPS 2026 · similarity 0.82)
- Intelligence per Watt: Measuring Intelligence Efficiency of Local AI (NeurIPS 2026 · similarity 0.81)
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.
