针对模拟存内计算噪声，提出自适应精度量化框架，显著提升低比特模型性能。
NANQ: Noise-Floor-Aware Mixed-Precision Non-Uniform Quantization for Analog Compute-in-Memory
- 基于实测噪声建模，动态分配不同精度，避开噪声主导区域的无效高精度
- 2比特权值量化下，视觉模型准确率提升8.05个百分点，语言模型困惑度降低54.7%
- 仅用3.2-3.8等效比特即获得接近全精度的收益，适合资源受限的边缘部署
模拟存内计算（CIM）可实现高效神经网络推理，但器件变异和读取噪声会严重损害低比特量化模型性能。现有面向CIM的量化方法主要最小化理想量化误差，忽视硬件噪声底限，导致精度分配低效。本文提出NANQ，一种面向模拟CIM的噪声感知混合精度非均匀量化框架。NANQ基于eFlash CIM阵列实测响应，建模权重幅度相关的噪声特性，并将噪声分布转化为自适应量化密度：在低噪声区域分配更细分辨率，在噪声主导区域避免无效高精度。进一步通过统一阈值识别各层在硬件噪声下的精度饱和点，实现层间位宽分配。片上实验在eFlash CIM SoC上验证，2比特权值幅度量化下，相较于PowerQuant，NANQ使视觉模型准确率提升8.05个百分点，语言模型平均困惑度（PPL）降低54.7%。混合精度的NANQ仅需3.2-3.8等效比特，即可捕获绝大部分额外量化资源带来的增益。
原文摘要 · Abstract (English)
Analog compute-in-memory (CIM) enables energy-efficient neural network inference, but device variation and read noise can severely degrade low-bit quantized models. Existing CIM-oriented quantization methods mainly minimize ideal quantization error, ignoring the hardware noise floor and thus causing inefficient precision allocation. We propose NANQ, a noise-aware mixed-precision non-uniform quantization framework for analog CIM. NANQ models magnitude-dependent weight noise from measured responses of an eFlash CIM array and converts the noise profile into an adaptive quantization density, assigning finer resolution to low-noise regions while avoiding ineffective precision in noise-dominated regions. It further assigns layer-wise bit-widths by identifying each layer's precision saturation point under hardware noise using a unified threshold. On-chip experiments on an eFlash CIM SoC show that, under 2-bit weight-magnitude quantization, NANQ improves vision-model accuracy by 8.05 percentage points and reduces language-model PPL by 54.7% on average over PowerQuant. Mixed-precision NANQ captures most of the gains obtainable from additional quantization resources with only 3.2-3.8 equivalent bits.
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。