Directed Graph Topology Inference via Graph Filter Identification
2026-06-29T07:23:58Z•8f18318bb962599765834a955efe8276f82e570572913deb9d97b77835b8c21e
Fokker--PlanckStiefel manifoldVC dimensionadversarial contaminationbenchmarkingcovariance estimationdataset selectionexact samplingglobal minimum-variance portfoliograph filtersgraph topology inferenceheat-ball representationsheavy tailsidentifiability','disentanglement'iterative hard thresholdingkappa-measurelatent SDEsminimax optimalitymodel rankingmonotonicitynon-affine aggregationpositive-only learningrobust regressionscore-based generative modelsstability
What happened
Collection of stat/ML papers (arXiv, 29 Jun 2026) covering theory and algorithms across graph identification, robust estimation, score-based generative modeling, learning-theory characterizations, dataset selection, and scaling laws. Key contributions include: (1) directed graph topology inference via graph-filter identification and Stiefel-constrained quadratic optimization; (2) exact decision-focused analysis of global minimum-variance portfolio regret under heavy tails; (3) AC-IHT — a two-stage adversarial-contamination-resistant iterative hard-thresholding algorithm with minimax near-opt-
Why it matters
A reviewed impact interpretation has not been published for this record.
Evidence and limitations
- Source ID
- arxiv_stat_ml
- Record identifier
- 8f18318bb962599765834a955efe8276f82e570572913deb9d97b77835b8c21e
- Enrichment time
- 2026-06-29T07:23:58Z
- AI-assisted enrichment
- Yes
This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.