Research map / Point clouds / Point cloud classification
CO-PILOT: Dynamic Top-Down Point Cloud with Conditional Neighborhood Aggregation for Multi-Gigapixel Histopathology Image Representation
Ramin Nakhli, Allen Zhang, Ali Mirabadi, Katherine Rich, Maryam Asadi, Blake Gilks, Hossein Farahani, Ali Bashashati
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use CNA layer instead of GAT (ablation) | CO-PILOT · HGSOC 1 | C-Index: 0.530 -> 0.568 | Improved |
| use CNA layer instead of GIN (ablation) | CO-PILOT · HGSOC 2 | C-Index: 0.520 -> 0.558 | Improved |
| add conditional aggregation to CNA (ablation) | CO-PILOT · HGSOC 1 | P-value: 0.05 -> < 0.01 | Improved |
| add Waterfall attention top-down aggregation (versus none) (ablation) | CO-PILOT · HGSOC 1 | P-value: 0.27 -> < 0.01 | Improved |
| make Waterfall attention conditional on higher-level features (ablation) | CO-PILOT · HGSOC 1 | C-Index: 0.551 -> 0.568 | Improved |
| add positional encoding by summation instead of concatenation (ablation) | CO-PILOT · HGSOC 1 | RMST: 0.66 -> 2.29 | Improved |
| add random rotation augmentation of node positions (ablation) | CO-PILOT · HGSOC 1 | C-Index: 0.523 -> 0.568 | Improved |
| add positional encoding in all DNPU layers instead of the first only (ablation) | CO-PILOT · HGSOC 1 | C-Index: 0.568 -> 0.568 | No clear effect |
| increase number of nearest neighbors K beyond 9 (ablation) | CO-PILOT · HGSOC 1 | C-Index | Hurt |
| share weights across all DNPUs (ablation) | CO-PILOT · HGSOC 2 | P-value: < 0.01 -> 0.97 | Hurt |
Closest papers by meaning
- Interpretable Vision-Language Survival Analysis with Ordinal Inductive Bias for Computational Pathology (ICLR 2025 · similarity 0.88)
- xMIL: Insightful Explanations for Multiple Instance Learning in Histopathology (NeurIPS 2024 · similarity 0.87)
- Dynamic Entity-Masked Graph Diffusion Model for Histopathology Image Representation Learning (AAAI 2025 · similarity 0.87)
- STimage-1K4M: A histopathology image-gene expression dataset for spatial transcriptomics (NeurIPS 2024 · similarity 0.87)
- M2OST: Many-to-one Regression for Predicting Spatial Transcriptomics from Digital Pathology Images (AAAI 2025 · similarity 0.87)
- Leveraging Tumor Heterogeneity: Heterogeneous Graph Representation Learning for Cancer Survival Prediction in Whole Slide Images (NeurIPS 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
