Researchers have developed a novel three-stage pipeline to optimize neural architectures for edge AI deployment, focusing on the interplay between Neural Architecture Search (NAS) and post-training quantization (PTQ). The pipeline includes a hardware-agnostic surrogate frontend, a quantization bridge with filtering, and an evolutionary backend for hardware mapping. An empirical study analyzed how INT4 PTQ affects the NAS Pareto space, finding that an FP32 zero-shot surrogate can outperform a dedicated INT4-trained surrogate in covering the Pareto space. AI
IMPACT This research could lead to more efficient and deployable AI models on edge devices by improving the synergy between architecture search and quantization techniques.
RANK_REASON The cluster contains an academic paper detailing a new method for optimizing neural architectures for edge AI. [lever_c_demoted from research: ic=1 ai=1.0]
- arXiv
- CGRA4ML
- Edge artificial intelligence
- Eleftherios Mylonas
- Hugging Face
- Int4
- NAS-Bench-201
- Neural architecture search
- Post Training Quantization Preprocessing Method of Convolutional Neural Network via Outlier Removal
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
