提出新方法提升早退出神经网络的推理效率与准确性
Rethinking Calibration for Early-Exit Neural Networks
- 引入失败预测机制，综合考虑准确率与计算成本
- 在多个数据集上实现更优的性价比平衡，优于传统校准方法
- 适合追求高效推理的模型部署场景
早退出神经网络（EENNs）通过中间分类器在预测置信度足够时提前终止计算以加速推理。现有方法多依赖置信度阈值决定退出，普遍认为提升分类器校准性可改善性能。本文挑战这一假设，指出仅优化校准不足以发挥自适应计算优势。为此，提出早退出失败预测（EEFP），同时考量预测正确性与继续计算的成本。设计了一种轻量级改进流程，可直接替代传统校准。大量实验表明，该方法在多项指标上优于校准策略，且EEFP更能可靠反映整体性能。代码已公开于https://github.com/gmum/rethinking-calibration-for-eenns。
原文摘要 · Abstract (English)
Early-exit neural networks (EENNs) accelerate inference by allowing intermediate classifiers to stop computation once predictions are confident enough. Most methods rely on confidence thresholds for exiting, and consequently, improving classifier calibration is widely assumed to improve performance. In this work, we challenge this assumption and show that calibration alone is not sufficient for EENNs to exploit adaptive computation. To address this insufficiency, we introduce Early-Exit Failure Prediction (EEFP), which accounts for both prediction correctness and the cost of further computation. We also propose a lightweight, EEFP-motivated procedure to improve the intermediate classifiers, which can directly replace calibration in EENNs. Extensive experiments demonstrate that our approach achieves superior cost-accuracy trade-offs compared to calibration, and EEFP more reliably reflects overall EENN performance. Our code is available at https://github.com/gmum/rethinking-calibration-for-eenns.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。