提出ARQ框架，让量化模型既保持高精度又具备可证明的鲁棒性。
ARQ: A Mixed-Precision Quantization Framework for Accurate and Certifiably Robust DNNs
- 用强化学习搜索混合精度量化方案，兼顾准确与鲁棒。
- 在多个基准上达到原浮点模型性能，仅用1.5%指令数。
- 适合需要高可靠性的模型部署场景，如自动驾驶。
混合精度量化已成为优化深度神经网络（DNN）执行的重要技术。然而，由于可证明鲁棒性验证成本过高，现有方法极少将其纳入量化考量。本文提出ARQ，一种创新的混合精度量化框架，不仅能保留平滑分类器的干净准确率，还可维持其可证明鲁棒性。ARQ利用强化学习寻找兼具准确性和鲁棒性的量化策略，并高效结合随机平滑这一主流统计验证算法。在所有基准和输入扰动水平下，ARQ均优于多个先进量化方法。经ARQ量化的网络性能接近原始浮点模型，同时仅需1.5%的指令数并实现最高认证半径。代码已开源：https://github.com/uiuc-arc/ARQ。
原文摘要 · Abstract (English)
Mixed precision quantization has become an important technique for optimizing the execution of deep neural networks (DNNs). Certified robustness, which provides provable guarantees about a model's ability to withstand different adversarial perturbations, has rarely been addressed in quantization due to the unacceptably high cost of certifying robustness. This paper introduces ARQ, an innovative mixed-precision quantization method that not only preserves the clean accuracy of the smoothed classifiers, but also maintains their certified robustness. ARQ uses reinforcement learning to find accurate and robust DNN quantization, while efficiently leveraging randomized smoothing, a popular class of statistical DNN verification algorithms. ARQ consistently performs better than multiple state-of-the-art quantization techniques across all the benchmarks and the input perturbation levels. The performance of ARQ quantized networks reaches that of the original DNN with floating-point weights, while using only 1.5% instructions and the highest certified radius. ARQ's code is available at https://github.com/uiuc-arc/ARQ.
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。