A Configuration-First Framework for Reproducible, Low-Code Localization

TL;DR

The paper addresses the reproducibility crisis in ML-driven radio localization by introducing LOCALIZE, a configuration-first, low-code framework that enforces reproducibility by default while remaining extensible. It articulates a three-component architecture (configuration, workflow orchestration, and artifacts) and a lean three-stage experiment pipeline (data preparation, training/optimization, evaluation/reporting) to enable rapid, fair comparisons. Through cross-platform qualitative evaluation against GUI/code-centric tools and a head-to-head notebook baseline, LOCALIZE demonstrates reduced authoring effort with comparable runtime and memory, and scalability up to 10x data without overhead explosion. The work also provides a reference implementation using DVC and Git, preconfigured datasets, and a clear path for future enhancements such as modality support and backend pluggability, highlighting practical impact for reproducible ML research in localization. Overall, LOCALIZE offers a practical blueprint for reproducible-by-design research environments that can be adapted to other scientific domains.

Abstract

Machine learning is increasingly permeating radio-based localization services. To keep results credible and comparable, everyday workflows should make rigorous experiment specification and exact repeatability the default, without blocking advanced experimentation. However, in practice, researchers face a three-way gap that could be filled by a framework that offers (i) low coding effort for end-to-end studies, (ii) reproducibility by default, including versioned code, data, and configurations, controlled randomness, isolated runs, and recorded artifacts, and (iii) built-in extensibility so new models, metrics, and stages can be added with minimal integration effort. Existing tools rarely deliver all three for machine learning in general and localization workflows in particular. In this paper, we introduce LOCALIZE, a low-code, configuration-first framework for radio localization in which experiments are declared in human-readable configuration files, a workflow orchestrator executes standardized pipelines from data preparation to reporting, and all artifacts, such as datasets, models, metrics, and reports, are versioned. Preconfigured, versioned datasets reduce initial setup effort and boilerplate, thereby accelerating model development and evaluation. The design, with explicit extension points, allows experts to add components without reworking the underlying infrastructure. Through a qualitative comparison and a head-to-head study against a plain Jupyter notebook baseline, we show that the framework reduces authoring effort while maintaining comparable runtime and memory behavior. Furthermore, using a Bluetooth Low Energy dataset, we demonstrate that scaling the training data from 1x to 10x keeps orchestration overheads bounded as data grows. Overall, the framework makes reproducible machine-learning-based localization experimentation practical, accessible, and extensible.

Figures (7)

• Figure 1: Outline of the proposed architecture consisting of three components: configuration, artifacts, workflow orchestrator.
• Figure 2: User workflow supported by the framework, separated by experience level. Novice users modify configurations, intermediate users adapt experiments or add datasets, expert users implement new stages or models.
• Figure 3: Lines of code added (positive) and removed (negative) when introducing successive changes in the Framework and Notebook implementations.
• Figure 4: Wall-clock and CPU time per pipeline stage for LOCALIZE and the Notebook implementation.
• Figure 5: Memory usage statistics per pipeline stage for LOCALIZE and the Notebook implementation. Bars show means with first/third quartiles as error bars.