AI4CITY - An Automated Machine Learning Platform for Smart Cities

Aug 4, 2026

Growing interest in machine learning solutions is creating demand for faster, less code-intensive ways to design and deploy models, yet many approaches require substantial expertise and large scripting efforts. This paper presents AI4CITY, an automated technological platform focused on smart cities applications. AI4CITY reduces ML design complexity via default end-to-end pipelines and a step-by-step workflow that includes task detection, data preprocessing, supervised learning model and hyperparameter selection, and deployment. Results are compared with popular AutoML tools such as H2O and AutoGluon, showing competitive performance.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages4

Reading Time10 min

AccessFree to download

Introduction

Automated ML motivation and expert shortage

AI4CITY platform overview and supported supervised tasks

Platform Architecture

CityCatalyst project context and agnostic ML goal

Default pipeline construction and required inputs

Expert customization and end-to-end deployment

Implementation approach with Python and imblearn logic

AI4CITY-An Automated Machine Learning Platform for Smart Cities Pedro José Pereira EPMQ-IT Engineering Maturity and Quality Lab, CCG ZGDV Institute ALGORITMI Centre/LASI University of Minho Guimarães, Portugal [pedro.pereira@ccg.pt](pedro.pereira@ccg.pt) Carlos Gonçalves EPMQ-IT Engineering Maturity and Quality Lab, CCG ZGDV Institute University of Minho Guimarães, Portugal [carlos.goncalves@ccg.pt](carlos.goncalves@ccg.pt) Lara Lopes Nunes NOS SGPS, S.A Lisboa, Portugal [lara.nunes@parceiros.nos.pt](lara.nunes@parceiros.nos.pt) Paulo Cortez ALGORITMI Centre/LASIDep. Information Systems University of Minho Guimarães, Portugal [pcortez@dsi.uminho.pt](pcortez@dsi.uminho.pt) André Pilastri EPMQ-IT Engineering Maturity and Quality Lab, CCG ZGDV Institute Guimarães, Portugal [andre.pilastri@ccg.pt](andre.pilastri@ccg.pt) ABSTRACT Nowadays, the general interest in Machine Learning (ML) based solutions is increasing. However, to develop and deploy a ML solution often requires experience and it involves developing large code scripts. In this paper, we propose AI4CITY, an automated technological platform that aims to reduce the complexity of designing ML solutions, with a particular focus on Smart Cities applications. We compare our solution with popular Automated ML (AutoML) tools (e.g., H2O, AutoGluon) and the results achieved by AI4CITY were quite interesting and competitive. CCS CONCEPTS • Computing methodologies → Learning settings; Machine learning algorithms; KEYWORDS Automated Machine Learning, Smart Cities, Supervised Learning ACM Reference Format: Pedro José Pereira, Carlos Gonçalves, Lara Lopes Nunes, Paulo Cortez, and André Pilastri. 2023. AI4CITY-An Automated Machine Learning Platform for Smart Cities. In The 38th ACM/SIGAPP Symposium on Applied Computing (SAC’23), March 27-March 31, 2023, Tallinn, Estonia. ACM, New York, NY, USA, Article 4, 4 pages. [https://doi.org/10.1145/3555776.3578740](https://doi.org/10.1145/3555776.3578740) 1 INTRODUCTION Due to advances in Information Technology (IT), nowadays it is more easy to collect, store and process data that reflects multiple aspects of our daily lives, giving rise to the concept of Smart Cities[4] . Permission to make digital or hard copies of part or all ofthis work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the owner/author(s) . SAC’23, March 27-March 31, 2023, Tallinn, Estonia © 2023 Copyright held by the owner/author(s) . ACM ISBN 978-1-4503-9517-5/23/03 . [https://doi.org/10.1145/3555776.3578740](https://doi.org/10.1145/3555776.3578740) All this data hold potentially valuable knowledge that can be extracted to better support decision-making. Thus, there has been a growing need to rely on data-driven systems, such as Artificial Intelligence (AI) and Machine Learning (ML) tools to make sense of the constant inputted data. As ML becomes so vast, in a way that is hard to keep up with, there is a shortage of experts that can effectively take advantage of state-of-art ML solutions. Thus, there has been a growing focus on the development of Automated ML (AutoML) solutions, which automate the search for the best ML algorithm and its hyperparameter setup [2, 9] . The paper proposes the AI4CITY technological platform, consisting of an AutoML tool that facilitates the application of ML algorithms to solve smart cities tasks. The goal is to reduce the complexity of the ML design code for smart cities applications, allowing both non-expert users and expert users to benefit from the whole ML workflow by using just a few lines of code. Our platform works with supervised learning tasks (classification, regression, and time series forecasting). In particular, it assumes a

FAQs

What problem does AI4CITY aim to solve in smart cities machine learning projects?

AI4CITY targets the complexity of designing and deploying ML solutions by reducing the amount of code and expertise required, especially for smart cities predictive analytics tasks.

Which machine learning tasks and data types does AI4CITY support?

AI4CITY focuses on supervised learning tasks, including classification, regression, and time series forecasting, and it targets tabular data in smart cities scenarios.

How does AI4CITY build and run its machine learning workflow?

It constructs default full ML pipelines that require input data and a target column, then automatically detects the ML task, applies necessary preprocessing, fits an AutoML algorithm, and outputs a ready-to-use pipeline for new data.

How does AI4CITY compare with other AutoML tools?

The paper compares AI4CITY with tools such as H2O and AutoGluon and reports that the achieved results are interesting and competitive.

Aug 4, 2026

Upload by Olivia Brown