Node-wise Feature Encoding for Neural Performance Prediction

2026-08-27

AI Summary

为解决边缘设备上神经网络性能预测问题，提出FeatureFormer，通过在门控图注意力架构中加入节点级计算成本编码来提高预测准确性。

Abstract

As neural networks are increasingly deployed on resource constrained edge devices, accurate prediction of latency and energy is critical for efficient neural architecture search. Existing GNN and transformer based predictors achieve strong results but largely ignore node-level computational cost, limiting their ability to model performance critical operations. To address this, we introduce FeatureFormer, a neural performance predictor that incorporates explicit node-wise encodings of FLOPs, parameter counts, and memory proxies within a gated graph attention architecture. We also present NNEQ, a new large-scale energy consumption dataset that enables unified evaluation of latency and energy prediction. Extensive experiments demonstrate that FeatureFormer achieves state-of-the-art performance across both metrics, including challenging out-of-domain settings. Finally, we show that the proposed encoding is broadly applicable and consistently improves existing predictors with negligible overhead.

Problem

• neural networks
• resource constrained edge devices
• latency and energy prediction
• node-level computational cost
• neural architecture search

Innovation

• node-wise feature encoding
• gated graph attention
• NNEQ dataset