Active Sampling: A Machine-Learning-Assisted Framework for Finite Population Inference with Optimal Subsamples

Data subsampling has become widely recognized as a tool to overcome computational and economic bottlenecks in analyzing massive datasets. We contribute to the development of adaptive design for estimation of finite population characteristics, using active learning and adaptive importance sampling. W...

Full description

Saved in:
Bibliographic Details
Published in:Technometrics Vol. 67; no. 1; pp. 46 - 57
Main Authors: Imberg, Henrik, Yang, Xiaomi, Flannagan, Carol, Bärgman, Jonas
Format: Journal Article
Language:English
Published: Alexandria Taylor & Francis 02.01.2025
American Society for Quality
Subjects:
Active learning
Adaptive importance sampling
Adaptive sampling
Adaptiveimportance sampling
Advanced driver assistance systems
Bayesian analysis
Computer and Information Sciences
Computer simulation experiments
Computer simulationexperiments
Data collection
Data- och informationsvetenskap (Datateknik)
Estimating techniques
Importance sampling
Inverse probability weighting
Inverseprobability weighting
Machine learning
Massive data points
Optimal design
Sampling methods
Simulation
Software
Traffic safety assessment
Traffic safetyassessment
ISSN:0040-1706, 1537-2723, 1537-2723
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Data subsampling has become widely recognized as a tool to overcome computational and economic bottlenecks in analyzing massive datasets. We contribute to the development of adaptive design for estimation of finite population characteristics, using active learning and adaptive importance sampling. We propose an active sampling strategy that iterates between estimation and data collection with optimal subsamples, guided by machine learning predictions on yet unseen data. The method is illustrated on virtual simulation-based safety assessment of advanced driver assistance systems. Substantial performance improvements are demonstrated compared to traditional sampling methods.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0040-1706
1537-2723
1537-2723
DOI:10.1080/00401706.2024.2374554