4-bit量化让机器人感知在分布外时失效，新方法用小模型修复且不增加算力。
Recti-Q: Feature-Space Rectification for Out-of-Distribution-Robust Quantized Perception in Edge Robotics
- 冻结量化模型，用少量源数据训练小型LoRA适配器修正特征空间。
- 4比特量化下在ImageNet-C和PACS上恢复近90%的分布外鲁棒性，部分超浮点精度。
- 参数增加不足1%，可低带宽远程更新，适合边缘机器人部署。
机器人感知系统越来越多依赖大视觉模型在资源受限的边缘设备上运行，后训练量化（PTQ）因能实现实时推理而备受青睐。然而，尽管PTQ通常保持干净的分布内准确率，我们发现其在实际部署中遭遇分布偏移（如传感器噪声、恶劣天气、新操作环境）时可靠性显著下降，形成‘量化导致的鲁棒性缺口’。在基础视觉基准（ImageNet-C和PACS）上，4比特PTQ模型虽分布内准确率损失微小，但分布外鲁棒性严重退化。为此，我们提出Recti-Q，一种轻量级特征空间修正框架：冻结量化骨干网络，仅用源数据训练一个小型分类头LoRA适配器。Recti-Q对CNN与Transformer均兼容，支持无需教师模型的高效训练，能恢复大量丢失的鲁棒性，在某些情况下甚至达到或超过浮点32位（FP32）性能。其参数开销低于1%（最小仅6KB），保留了超过99%的量化内存节省，计算开销可忽略，并支持低带宽过空中（OTA）远程补丁，适用于在不可预测物理环境中运行的机器人集群。
原文摘要 · Abstract (English)
Robotic perception pipelines increasingly rely on large vision backbones deployed on SWaP-constrained edge platforms, making post-training quantization (PTQ) attractive for real-time inference. However, while PTQ often preserves clean in-distribution accuracy, we show that it can substantially degrade reliability under deployment-relevant distribution shifts (e.g., sensor noise, severe weather, and novel operating environments), creating a Quantization-Induced Robustness Gap. Across foundational vision benchmarks (ImageNet-C and PACS), 4-bit PTQ models exhibit pronounced robustness degradation despite negligible ID accuracy loss. To address this, we propose Recti-Q, a lightweight feature-space rectification framework that freezes the quantized backbone and trains a small classifier-head LoRA adapter using only source data. Recti-Q is architecture-agnostic across CNNs and Transformers, supports efficient teacher-free training, and recovers a significant portion of the lost robustness, in some cases matching or exceeding FP32 performance. At less than 1% parameter overhead (as small as 6 KB), Recti-Q preserves over 99% of PTQ memory savings, adds negligible compute, and enables low-bandwidth Over-The-Air (OTA) resilience patching for deployed robotic fleets operating in unpredictable physical environments.
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。