# Loss Distillation via Gradient Matching for Point Cloud Completion with Weighted Chamfer Distance

arXiv: https://arxiv.org/abs/2409.06171 · HTML: https://mlsanity.com/paper/2409.06171/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 33 of its 51 references that are in the map.
- First posted on arXiv 10 Sep 2024.

## Builds on (most cited first, 30 of 33)

- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [ShapeNet: An Information-Rich 3D Model Repository](https://mlsanity.com/paper/1512.03012/) (2015)
- [Knowledge Distillation: A Survey](https://mlsanity.com/paper/2006.05525/) (2020)
- [Accelerating 3D Deep Learning with PyTorch3D](https://mlsanity.com/paper/2007.08501/) (2020)
- [Deep Learning for 3D Point Clouds: A Survey](https://mlsanity.com/paper/1912.12033/) (2019)
- [AtlasNet: A Papier-Mach'e Approach to Learning 3D Surface Generation](https://mlsanity.com/paper/1802.05384/) (2018)
- [PCN: Point Completion Network](https://mlsanity.com/paper/1808.00671/) (2018)
- [Knowledge Distillation and Student-Teacher Learning for Visual Intelligence: A Review and New Outlooks](https://mlsanity.com/paper/2004.05937/) (2020)
- [PoinTr: Diverse Point Cloud Completion with Geometry-Aware Transformers](https://mlsanity.com/paper/2108.08839/) (2021)
- [What Do Single-view 3D Reconstruction Networks Learn?](https://mlsanity.com/paper/1905.03678/) (2019)
- [Bilevel Optimization: Convergence Analysis and Enhanced Design](https://mlsanity.com/paper/2010.07962/) (2020)
- [GRNet: Gridding Residual Network for Dense Point Cloud Completion](https://mlsanity.com/paper/2006.03761/) (2020)
- [SnowflakeNet: Point Cloud Completion by Snowflake Point Deconvolution with Skip-Transformer](https://mlsanity.com/paper/2108.04444/) (2021)
- [Cascaded Refinement Network for Point Cloud Completion](https://mlsanity.com/paper/2004.03327/) (2020)
- [Density-aware Chamfer Distance as a Comprehensive Metric for Point Cloud Completion](https://mlsanity.com/paper/2111.12702/) (2021)
- [Grasp Pose Detection in Point Clouds](https://mlsanity.com/paper/1706.09911/) (2017)
- [A Conditional Point Diffusion-Refinement Paradigm for 3D Point Cloud Completion](https://mlsanity.com/paper/2112.03530/) (2021)
- [PMP-Net: Point Cloud Completion by Learning Multi-step Point Moving Paths](https://mlsanity.com/paper/2012.03408/) (2020)
- [SeedFormer: Patch Seeds based Point Cloud Completion with Upsample Transformer](https://mlsanity.com/paper/2207.10315/) (2022)
- [Detail Preserved Point Cloud Completion via Separated Feature Aggregation](https://mlsanity.com/paper/2007.02374/) (2020)
- [PointAttN: You Only Need Attention for Point Cloud Completion](https://mlsanity.com/paper/2203.08485/) (2022)
- [Snowflake Point Deconvolution for Point Cloud Completion and Generation with Skip-Transformer](https://mlsanity.com/paper/2202.09367/) (2022)
- [Edge Grasp Network: A Graph-Based SE(3)-invariant Approach to Grasp Detection](https://mlsanity.com/paper/2211.00191/) (2022)
- [Knowledge Distillation in Deep Learning and its Applications](https://mlsanity.com/paper/2007.09029/) (2020)
- [LAKe-Net: Topology-Aware Point Cloud Completion by Localizing Aligned Keypoints](https://mlsanity.com/paper/2203.16771/) (2022)
- [3D Local Features for Direct Pairwise Registration](https://mlsanity.com/paper/1904.04281/) (2019)
- [DPDist : Comparing Point Clouds Using Deep Point Cloud Distance](https://mlsanity.com/paper/2004.11784/) (2020)
- [FBNet: Feedback Network for Point Cloud Completion](https://mlsanity.com/paper/2210.03974/) (2022)
- [Imagination Policy: Using Generative Point Cloud Models for Learning Manipulation Policies](https://mlsanity.com/paper/2406.11740/) (2024)
- [OrbitGrasp: SE(3)-Equivariant Grasp Learning](https://mlsanity.com/paper/2407.03531/) (2024)

## Built on by (most cited first, 4 of 4)

- [DINE: Distance Is Not Enough -- Learning Global Deformation Priors for Robust Soft-Tissue Point Cloud Registration](https://mlsanity.com/paper/2607.14946/) (2026)
- [Med-PU: Point Cloud Upsampling for High-Fidelity 3D Medical Shape Reconstruction](https://mlsanity.com/paper/2501.16716/) (2025)
- [A Strong View-Free Baseline Approach for Single-View Image Guided Point Cloud Completion](https://mlsanity.com/paper/2506.15747/) (2025)
- [NexusFlow: Unifying Disparate Tasks under Partial Supervision via Invertible Flow Networks](https://mlsanity.com/paper/2512.06251/) (2025)

## Abstract (arXiv)

3D point clouds enhanced the robot's ability to perceive the geometrical information of the environments, making it possible for many downstream tasks such as grasp pose detection and scene understanding. The performance of these tasks, though, heavily relies on the quality of data input, as incomplete can lead to poor results and failure cases. Recent training loss functions designed for deep learning-based point cloud completion, such as Chamfer distance (CD) and its variants (eg HyperCD ), imply a good gradient weighting scheme can significantly boost performance. However, these CD-based loss functions usually require data-related parameter tuning, which can be time-consuming for data-extensive tasks. To address this issue, we aim to find a family of weighted training losses (em weighted CD) that requires no parameter tuning. To this end, we propose a search scheme, em Loss Distillation via Gradient Matching, to find good candidate loss functions by mimicking the learning behavior in backpropagation between HyperCD and weighted CD. Once this is done, we propose a novel bilevel optimization formula to train the backbone network based on the weighted CD loss. We observe that: (1) with proper weighted functions, the weighted CD can always achieve similar performance to HyperCD, and (2) the Landau weighted CD, namely em Landau CD, can outperform HyperCD for point cloud completion and lead to new state-of-the-art results on several benchmark datasets. it Our demo code is available at urlhttps://github.com/Zhang-VISLab/IROS2024-LossDistillationWeightedCD.