Research map / Graph learning / GNN architectures
Graph Coloring for Multi-Task Learning
Santosh Patapati, Ian Noronha
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 |
|---|---|---|---|
| add dynamic recoloring instead of static one-shot coloring (ablation) | SON-GOKU · CIFAR-10 | accuracy (%): 61 -> 65 | Improved |
| add dynamic recoloring instead of static one-shot coloring (ablation) | SON-GOKU · NYUv2 | depth RMSE: 0.65 -> 0.59 | Improved |
| add history-averaged conflict estimates instead of single-step (H=1) (ablation) | SON-GOKU · CIFAR-10 | accuracy (%): 40 -> 65 | Improved |
| add history-averaged conflict estimates instead of single-step (H=1) (ablation) | SON-GOKU · MM-IMDb | accuracy (%): 20 -> 61 | Improved |
| use thresholded conflict graph instead of signed-only graph (ablation) | SON-GOKU · CIFAR-10 | accuracy (%): 56 -> 65 | Improved |
| use thresholded conflict graph instead of kNN-symmetric graph (ablation) | SON-GOKU · CIFAR-10 | accuracy (%): 60 -> 65 | Improved |
| use quantile threshold graph instead of fixed threshold graph (density-matched) (ablation) | SON-GOKU · NYUv2 | depth RMSE: 0.59 -> 0.59 | No clear effect |
| add SON-GOKU scheduling on top of AdaTask | AdaTask · CIFAR-10 | accuracy (%): 63 -> 67 | Improved |
| add SON-GOKU scheduling on top of AdaTask | AdaTask · MM-IMDb | accuracy (%): 59 -> 63 | Improved |
| add SON-GOKU scheduling on top of PCGrad | PCGrad · F&B | accuracy (%): 50 -> 54 | Improved |
| add SON-GOKU scheduling on top of PCGrad | PCGrad · HEALTH | accuracy (%): 58 -> 62 | Improved |
| add SON-GOKU scheduling with GradNorm warm start | GradNorm · NYUv2 | surface-normal angle error: 21.4 -> 19.6 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Uniform — UniFORM: Towards Unified Framework for Anomaly Detection on Graphs (AAAI 2025) · better on 10 of 10 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Selective Task Group Updates for Multi-Task Optimization (ICLR 2025 · similarity 0.87)
- DMTG: One-Shot Differentiable Multi-Task Grouping (ICML 2024 · similarity 0.86)
- Towards Principled Task Grouping for Multi-Task Learning (NeurIPS 2026 · similarity 0.85)
- Quantifying Task Priority for Multi-Task Optimization (CVPR 2024 · similarity 0.85)
- Decoupling Weighing and Selecting for Integrating Multiple Graph Pre-training Tasks (ICLR 2024 · similarity 0.84)
- Towards Consistent Multi-Task Learning: Unlocking the Potential of Task-Specific Parameters (CVPR 2025 · similarity 0.84)
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.
