Structure-Aware Graph Abstention for Reliable Selective Forecasting
We treat instance-level plausibility and relational consistency as distinct reliability axes and operationalize the latter via a learned sparse graph and a Dirichlet-style structural energy Estruct, trained with error-weighted graph regularization and score-error alignment.
ProofPaper ↗
Key points
- Selective forecasting abstains on high-risk test windows under a retained-coverage budget.
- Existing gates such as TEM (Brusokas et al., 2025) score each forecast as a whole; for multivariate outputs, trajectories can look plausible while violating dependencies among variables.
- On seven long-horizon benchmarks and four backbones, structural gating often reduces selective MSE versus TEM at matched coverage, with the largest gains where cross-variable structure appears more informative in our benchmarks; gains are not universal, indicating a complementary abstention signal.
- Table 1 is a Protocol A ranking diagnostic (seed 2024); three-seed deployable Protocol B on an aligned subset is in Table 3 (full validation-to-test grids: Appendix A).
Sources (1)
- [1]Structure-Aware Graph Abstention for Reliable Selective ForecastingarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 01:23 PM
We treat instance-level plausibility and relational consistency as distinct reliability axes and operationalize the latter via a learned sparse graph and a Dirichlet-style structural energy E_struct, trained with error-weighted graph regularization and score-error alignment.
Selective forecasting abstains on high-risk test windows under a retained-coverage budget.
Extractive summary: sentences quoted from the sources.