对比联邦学习与知识蒸馏在点云分类中的表现，发现联合使用可提升模型效率但需注意评估陷阱。
Benchmarking Federated Learning and Knowledge Distillation for Point Cloud Classification
- 构建130种组合的基准测试，系统评估联邦学习与知识蒸馏在点云分类中的效果。
- 蒸馏后模型体积缩小74.51%，推理速度提升一倍，性能接近甚至超过教师模型。
- 建议使用无标签蒸馏避免评估偏差，确保结果反映联邦学习的真实性能。
在隐私敏感且资源受限的场景中部署3D点云分析面临两大挑战：数据无法集中，模型需运行于边缘设备。本文提出一个跨13种联邦学习（FL）算法与10种知识蒸馏（KD）目标（共130种组合）的多种子基准测试，涵盖504次训练，在ModelNet40和临床颅缝早闭数据集上评估。结果显示：极端非独立同分布（non-IID）标签偏移下，纯联邦学习性能显著下降——ModelNet40上最优方法仅达76.32%（集中式参考为92.26%），临床数据上最高75.83%（集中式为100%）。知识蒸馏可将教师模型压缩至原体积的25.49%，推理速度提升约一倍，且学生模型常超越或匹配教师。然而联合流程存在评估陷阱：若蒸馏保留标签代理集上的硬标签交叉熵项，即使联邦教师模型崩溃（准确率仅8.50%），学生仍可达92.94%。该84.4分差距源于代理标签而非真实联邦模型，违背了联邦学习的隐私初衷。不含硬标签的目标能准确追踪教师质量（相关系数r≈0.99），当教师失效时亦会同步退化。因此建议采用无标签蒸馏评估，以确保报告准确率真正反映联邦教师性能。
原文摘要 · Abstract (English)
Deploying 3D point cloud analysis in privacy-sensitive, resource-constrained settings faces two barriers: data cannot be centralized, and models must run on limited edge hardware. We present a multi-seed benchmark jointly evaluating federated learning (FL) and knowledge distillation (KD) for 3D point cloud classification. It spans 13 FL algorithms and 10 KD objectives (a 130-pair cross-product) across 504 training runs, evaluated on ModelNet40 and a clinical craniosynostosis dataset. We report three findings. First, under extreme non-IID label skew, standalone FL degrades sharply: on ModelNet40, the strongest method reaches 76.32% against a 92.26% centralized reference; on clinical data, the best reaches 75.83% against 100%. Second, distillation successfully compresses the teacher into a student 74.51% smaller and roughly twice as fast at inference, often matching or surpassing the teacher. Third, the combined pipeline exposes an evaluation pitfall: when distillation keeps a hard-label cross-entropy term on a labeled proxy split, a collapsed federated teacher (8.50%) paired with Logit-MSE still yields a 92.94% student. This 84.4-point gap reflects the proxy labels rather than the federated model, reusing the very labels whose privacy motivated federation. Objectives without hard labels instead track teacher quality ($r \approx 0.99$) and collapse when the teacher does. We therefore recommend evaluating FL-KD pipelines with label-free distillation so reported accuracy reflects the federated teacher, not the proxy.
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。