Recent years have seen an accelerated interest in applying artificial intelligence (AI) and, in particular, machine learning (ML) techniques to optimize, design, and automate communication networks. As the complexity of modern 5G and beyond networks grows, ML techniques promise new design methodologies and tools for architecting complex communication networks and automating network operation. At the same time, the increasingly powerful and ubiquitous sensing, computation and communication capabilities of emerging networks are driving ML computations closer to the edge, enabling local processing of distributed datasets; a shift that promises pervasive and contextualized ML solutions at scale.

ML techniques are rapidly evolving, moving from traditional supervised learning to self-supervised learning, reinforcement learning, and artificial general intelligence. However, they face several challenges for wide-scale deployment in wireless systems. Wireless communication networks should operate with almost no access to labeled data and must be particularly robust to dynamically-changing network environments that vary at extremely short timescales. Suitable ML solutions should thus (i) generalize well to new scenarios with minimal supervision and (ii) deliver effective performance for mobile and severely resource-constrained devices. Addressing these challenges requires effectively combining wireless domain knowledge with a deep understanding of ML methods towards developing high-performance and practical solutions and optimizing the performance of existing and future communication systems.

Exploiting synergies between applying ML techniques to optimize communication networks and distributing ML workloads across resource-constrained and unreliable networks requires a cross-disciplinary virtuous co-design cycle that could transform both wireless networking and ML technologies. Modern communication networks need to support diverse and mission-critical services over time-varying wireless channels, mandating automated approaches that can operate online, reliably, and in real-time. At the same time, learning at the wireless edge must address several challenges, including device heterogeneity and resource constraints, privacy concerns, limited data, and partially-observable environments.

This workshop seeks to bring ML and wireless networking experts together to identify inter-disciplinary approaches to evolve ML algorithms for and over communication networks that operate under constrained resources, including time, labeling, and computational capacity constraints. The workshop will provide a unique opportunity to expose the MLSys community to the challenges and opportunities of integrating ML methods into resource-constrained communication networks. It will also highlight emerging trends in ML with limited resources and their implications for the design and operation of next-generation communication networks.

We are seeking original submissions in topics including, but not limited to:

• Learning in wireless networks with limited training data
• Multi-agent federated/distributed learning with low computational and communication resources
• Communicated data compression for network-wide task completion
• Online learning with wireless latency constraints
• Learning in wireless networks with privacy constraints
• Few-shot learning and adaptation in wireless environments
• Datasets and benchmarks for resource-constrained learning in wireless networks

Key Dates:

• Paper Submission Open: February 13, 2023.
• Paper Submission Deadline: March 21, 2023 (EXTENDED), 11:59 pm Eastern Time.
• Acceptance Notification: May 1, 2023.
• Camera Ready Deadline: May 15, 2023, 11:59 pm Eastern Time.
• Workshop presentations: June 8, 2023.