Online Conformal Prediction: Enforcing monotonicity via Online Optimization
2026-05-14T07:23:58Z•3e203a3b0391161c3a2ae9336aa1a229d0756dfb0cb54d4485b847bc4094c50c
A/B-testingISOMORPHLLM-workflowsalways-valid-inference','e-process'arXivattention-scalingbanditsbenchmarksdigital-twinexperimental-designfeature-elicitationgenerate-verifyinverse-temperaturemachine-learningnested-prediction-setsonline-conformal-predictiononline-optimizationreward-modelingrobust-designself-attentionstatisticssupply-chaintime-series-forecastingtransformersweak-to-strong-generalization
What happened
Collection of new ML/statistics arXiv papers (May 14, 2026) covering methodological advances and benchmarks. Highlights: (1) Online conformal prediction methods that produce simultaneously valid, strictly nested prediction sets across coverage levels via online optimization. (2) A unified theory for inverse-temperature (logit rescaling) scaling in self-attention based on a gap-counting function, reconciling prior scaling laws. (3) ISOMORPH: an open supply-chain digital twin and dataset generator for multi-echelon logistics with interpretable, configurable parameters and benchmark rollouts. (4)
Why it matters
A reviewed impact interpretation has not been published for this record.
Evidence and limitations
- Source ID
- arxiv_stat_ml
- Record identifier
- 3e203a3b0391161c3a2ae9336aa1a229d0756dfb0cb54d4485b847bc4094c50c
- Enrichment time
- 2026-05-14T07:23:58Z
- 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.