mlsanityData to 14 Oct 2026

Purge-Gate: Backpropagation-Free Test-Time Adaptation for Point Clouds Classification via Token Purging

arXiv 2509.09785 ↗First posted 11 Sep 2025cs.CVcited by 3

Not on the map. Why?

Abstract (arXiv)

Test-time adaptation (TTA) is crucial for mitigating performance degradation caused by distribution shifts in 3D point cloud classification. In this work, we introduce Token Purging (PG), a novel backpropagation-free approach that removes tokens highly affected by domain shifts before they reach attention layers. Unlike existing TTA methods, PG operates at the token level, ensuring robust adaptation without iterative updates. We propose two variants: PG-SP, which leverages source statistics, and PG-SF, a fully source-free version relying on CLS-token-driven adaptation. Extensive evaluations on ModelNet40-C, ShapeNet-C, and ScanObjectNN-C demonstrate that PG-SP achieves an average of +10.3% higher accuracy than state-of-the-art backpropagation-free methods, while PG-SF sets new benchmarks for source-free adaptation. Moreover, PG is 12.4 times faster and 5.5 times more memory efficient than our baseline, making it suitable for real-world deployment. Code is available at hyperlinkhttps://github.com/MosyMosy/Purge-Gatehttps://github.com/MosyMosy/Purge-Gate

Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Builds on 37 of its 46 references that are in the map. First posted on arXiv 11 Sep 2025.

Builds on

37 of 46 references.

References · 37

  1. 2017Attention Is All You Needcited by 60,361
  2. 2015Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shiftcited by 7,005
  3. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,248
  4. 2023Mamba: Linear-Time Sequence Modeling with Selective State Spacescited by 4,718
  5. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,401
  6. 2015ShapeNet: An Information-Rich 3D Model Repositorycited by 2,897
  7. 2018Dynamic Graph CNN for Learning on Point Cloudscited by 2,223
  8. 2024VMamba: Visual State Space Modelcited by 969
  9. 2019Revisiting Point Cloud Classification: A New Benchmark Dataset and Classification Model on Real-World Datacited by 532
  10. 2022Masked Autoencoders for Point Cloud Self-supervised Learningcited by 459
  11. 2020Improving robustness against common corruptions by covariate shift adaptationcited by 431
  12. 2021MEMO: Test Time Robustness via Adaptation and Augmentationcited by 377
25 more
  1. 2022Point-M2AE: Multi-scale Masked Autoencoders for Hierarchical Point Cloud Pre-trainingcited by 264
  2. 2020Evaluating Prediction-Time Batch Normalization for Robustness under Covariate Shiftcited by 236
  3. 2022Test-Time Training with Masked Autoencoderscited by 212
  4. 2022Parameter-free Online Test-time Adaptationcited by 189
  5. 2024PointMamba: A Simple State Space Model for Point Cloud Analysiscited by 167
  6. 2022Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoderscited by 144
  7. 2021Learned Token Pruning for Transformerscited by 139
  8. 2023TTN: A Domain-Shift Aware Batch Normalization in Test-Time Adaptationcited by 101
  9. 2022Benchmarking Robustness of 3D Point Cloud Recognition Against Common Corruptionscited by 80
  10. 2022MM-TTA: Multi-Modal Test-Time Adaptation for 3D Semantic Segmentationcited by 76
  11. 2022Back to the Source: Diffusion-Driven Test-Time Adaptationcited by 71
  12. 2023Dynamic Token Pruning in Plain Vision Transformers for Semantic Segmentationcited by 62
  13. 2023PPT: Token Pruning and Pooling for Efficient Vision Transformerscited by 38
  14. 2023Continual Test-time Domain Adaptation via Dynamic Sample Selectioncited by 35
  15. 2024WATT: Weight Average Test-Time Adaptation of CLIPcited by 33
  16. 2022MATE: Masked Autoencoders are Online 3D Test-Time Learnerscited by 23
  17. 2024Backpropagation-free Network for 3D Test-time Adaptationcited by 14
  18. 2024CLIPArTT: Adaptation of CLIP to New Domains at Test Timecited by 14
  19. 2023Rapid Network Adaptation: Learning to Adapt Neural Networks Using Test-Time Feedbackcited by 10
  20. 2025Spectral Informed Mamba for Robust Point Cloud Processingcited by 10
  21. 2025Test-Time Adaptation of Vision-Language Models for Open-Vocabulary Semantic Segmentationcited by 10
  22. 2024Test-Time Adaptation in Point Clouds: Leveraging Sampling Variation with Weight Averagingcited by 7
  23. 2025Spectral State Space Model for Rotation-Invariant Visual Representation Learningcited by 4
  24. 2025SMART-PC: Skeletal Model Adaptation for Robust Test-Time Training in Point Cloudscited by 4
  25. 2024GeoMask3D: Geometrically Informed Mask Selection for Self-Supervised Point Cloud Learning in 3Dcited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 2.

3 papers.

  1. 2025TRUST: Test-Time Refinement using Uncertainty-Guided SSM Traversescited by 0
  2. 2026MAMVI: 3D Test-Time Adaptation via Masked Multi-View Point Cloudscited by 0
  3. 2026Test Time Adaptation Methods for Point Cloud Registration in Laparoscopic Surgerycited by 0