Welcome to AIoT 2025!

The rapid growth of connected devices—expected to exceed 30 billion by 2030—is reshaping how we collect and process information from the physical world. Devices ranging from smartphones and wearables to industrial sensors and microcontrollers are continuously generating vast amounts of data. Meanwhile, advanced wireless technologies such as 5G, Wi-Fi 6/7, and LPWAN are making it possible to connect a wide range of devices across large and diverse environments.

In this setting, emerging approaches like distributed, Federated, and Edge Learning are gaining momentum. These methods bring intelligence closer to the data, reducing communication overhead, enabling faster decisions, improving privacy, and supporting energy-efficient processing. Such techniques are particularly relevant in IoT systems that blend local sensing, computation, and actuation—encompassing not only traditional IoT deployments but also edge-cloud infrastructures, cyber-physical systems, and collaborative networks of autonomous agents.

At the same time, integrating Distributed Learning into these IoT systems introduces several open challenges. These include compressing models to transmit over constrained or unreliable networks, managing limited and heterogeneous resources at the edge, accelerating training under dynamic conditions, and ensuring security and privacy when data remains locally distributed.

The Third International Workshop on the Integration between Distributed Machine Learning and the Internet of Things (AIoT) serves as a venue for researchers and practitioners exploring how Distributed and Federated Learning can be applied in real-world, resource-constrained, and large-scale IoT systems. Topics span a wide range of connected IoT systems—where decentralized intelligence plays a key role in enabling robust, adaptive, and efficient operations.

Call for Papers

The rapid growth of connected devices—projected to surpass 30 billion by 2030—is transforming how data is collected, processed, and acted upon in real-world environments. In this evolving landscape, emerging paradigms like Distributed, Federated, and Edge Learning are becoming essential. These approaches bring intelligence closer to the data source, minimizing communication overhead, reducing latency, improving privacy, and enabling energy-efficient computation. However, integrating Distributed Learning into IoT systems presents several open challenges: how to compress models for constrained networks, how to manage limited and heterogeneous edge resources, how to accelerate training under dynamic conditions, and how to preserve security and privacy while keeping data decentralized, to name a few.

AIoT, the Third International ACM MobiHoc Workshop on the Integration between Distributed Machine Learning and the Internet of Things, aims to bring together researchers and practitioners from academia and industry to explore the design, deployment, and operation of distributed intelligence in resource-constrained and large-scale IoT systems.

We invite original contributions in the form of theoretical insights, algorithmic advances, experimental evaluations, and real-world applications. Topics of interest include, but are not limited to:

• Efficient Machine Learning on low-power or constrained IoT systems

• Distributed, Federated, and Split Learning across edge and cloud systems in IoT environments

• System architectures and runtime optimization for learning in IoT systems

• Hardware acceleration and platform co-design for edge intelligence in IoT systems

• Communication and networking support for distributed model training in IoT systems

• Protocols for model sharing, updates, and coordination in IoT systems

• Edge collaboration and cross-device intelligence in IoT systems Privacy-preserving training methods and secure aggregation mechanisms for distributed, Federated, and Edge Learning in IoT systems

• Experimental testbeds, real-world deployments, and benchmarking tools for IoT systems

• Applications in areas such as smart cities, healthcare, industrial IoT systems, agriculture, and transportation

• Scalability, reliability, and performance tuning for large-scale IoT systems

• Open challenges, new directions, and emerging trends in decentralized learning for IoT systems

• Model personalization and adaptation techniques for Federated Learning in IoT systems

• Fault tolerance, robustness, and reliability in Distributed Learning for IoT systems

• Edge AI for low-latency applications in IoT systems

• Energy-aware learning algorithms for IoT systems

• Cross-platform Machine Learning for heterogeneous IoT systems

• Network slicing and QoS-aware techniques for Federated Learning in IoT systems

• Evolutionary models and online learning techniques in IoT systems

• Decentralized consensus algorithms for model coordination in IoT systems

• AI techniques for IoT security

Important Dates

• Paper submission: July 30, 2025 (11:59pm EDT)

• Acceptance notification: August 23, 2025

• Camera ready: August 30, 2025

• Workshop: October 30, 2025

Submission Instructions

Papers should be submitted via the HotCRP submissing website (coming soon).

Submissions must be original, unpublished work, and not currently under consideration elsewhere. Papers should not exceed 6 pages (US letter size) double column including figures, tables, and references in standard ACM format. Papers must be submitted electronically in printable PDF form. Templates for the standard ACM format can be found at this link: http://www.acm.org/publications/article-templates/proceedings-template.html . If you are using LaTeX, please refer to the sample file “sample-sigconf.tex” after you download the .zip templates file and unzip it. Note that the document class “\documentclass[sigconf]{acmart}” should be used. No changes to margins, spacing, or font sizes are allowed from those specified by the style files. Papers violating the formatting guidelines will be returned without review.

All submissions will be reviewed using a single-blind review process. The identity of referees will not be revealed to authors, but author can keep their names on the submitted papers, on figures, bibliography, etc.

Dual Submission Policy

Accepted papers will appear in the conference proceedings published by the ACM. Warning: It is ACM policy not to allow double submissions, where the same paper is submitted to more than one conference/journal concurrently. Any double submissions detected will be immediately rejected from all conferences/journals involved.