Get Started
Home
Topics
Search
Library
Research questionHow can counterfactual KPI prediction retain finite-sample coverage under hidden confounding with scarce randomized telemetry?Network telemetry can hide variables that influence control actions, making counterfactual KPI estimates unreliable. Randomized telemetry can protect validity, but its scarcity may make the resulting prediction sets too broad to be useful.
Machine Learning
Research Paper
Statistical Machine Learning
Technology
Latest papersRecent research connected to this question, newest first.Confounding-Valid Conformal Inference for Counterfactual KPIs in Wireless NetworksThe setting is wireless-network and radio access network control, with abundant potentially confounded observational telemetry and limited randomized telemetry. The evidence covers finite-sample coverage under arbitrary hidden confounding and prediction-set efficiency in two representative RAN control tasks.research paper · Sep 4, 2026
Related questions
How can we estimate counterfactual outcomes with valid uncertainty as treatment patterns shift?How can predictive ML inference share wireless access points without degrading packet services under load?How can lightweight time-series forecasters learn accurate, specialized forecasts from scarce private data?How can we reliably select a heterogeneous treatment effect estimator without observing treatment effects?