对比迭代与加宽，发现不同场景下哪种方法更有效。
Iterate or Widen? When Test-Time Refinement Helps LiDAR Scene Completion: A Controlled Study of Evidence Geometry, Training Coverage, and Compute
- 用固定模型迭代优化或扩展网络宽度，比较测试时性能提升。
- 连续角度缺失下迭代提升mIoU 0.911点，优于加宽模型。
- 稀疏数据靠训练覆盖解决，噪声干扰需新方法应对。
在激光雷达语义场景补全任务中，研究模型应通过迭代优化使用额外测试计算，还是用相同参数预算构建更宽的一次性预测器。通过控制协议分离出连贯区域移除、独立稀疏化、距离依赖衰减和叠加杂波四种干扰因素，并保持精确的场景配对。在五个训练种子和815个SemanticKITTI序列-08帧上，完整迭代系统在连续角度缺失条件下比加宽对照组提升mIoU 0.911点（95%置信区间[0.804, 1.040]），超过预设0.5点实用阈值。在独立75%稀疏化下，迭代仅提升0.300点（[0.166, 0.436]），而观察族增强提升5.975点（[5.662, 6.140]）。两种干预均无法修复叠加杂波问题。迭代系统每帧耗时10.74毫秒、占用0.75 GiB内存，加宽模型为6.25毫秒、0.23 GiB。结果表明：几何条件决定最优策略——连贯空缺适合固定深度迭代，广泛稀疏依赖训练覆盖，虚假证据需另辟鲁棒机制。
原文摘要 · Abstract (English)
Should a completion model spend extra test-time compute by iterating, or spend a similar parameter budget on a wider one-shot predictor? The answer is easily confounded by denoising curricula, corruption augmentation, capacity, and unpaired evaluation. We study this question in LiDAR semantic scene completion by comparing a one-shot predictor, a parameter-matched wider predictor, and a weight-tied multigrid refiner initialized from the same frozen predictor. The protocol separates coherent region removal, independent thinning, range-dependent attenuation, and additive clutter while preserving exact scene-condition pairing. Across five training seeds and 815 SemanticKITTI sequence-08 frames, the full iterative system improves mIoU over the wide control by 0.911 points under contiguous angular removal, with a 95% moving-block bootstrap interval of [0.804, 1.040] that clears a predeclared 0.5-point practical margin. Under independent 75% thinning, iteration adds only 0.300 points [0.166, 0.436], whereas observation-family augmentation adds 5.975 points [5.662, 6.140]. Neither intervention repairs additive clutter. The iterative system also costs 10.74 ms and 0.75 GiB per frame, versus 6.25 ms and 0.23 GiB for the wide control. These results establish a geometry-conditioned empirical boundary rather than a universal advantage: coherent gaps can justify fixed-depth refinement, broadly thinned evidence is addressed more effectively by training coverage, and spurious evidence requires a different robustness mechanism.
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。