1st International Workshop on Programmable and Learning Architectures for Intelligent Networks (PLAIN)
Brief Description: The concepts of virtualization, softwarization, and programmability, introduced into wireless and op- tical networks, are now central to the transformation of communication infrastructures. These shifts enable the transition from static, hardware-based systems to dynamic, software-defined environments that can support the stringent requirements of emerging AI workloads. In this landscape, the network is no longer just a conduit for data, but a programmable and intelligent fabric that enables end-to-end AI functionality.
Hence, the integration of AI into network infrastructures demands a rethinking of architectures at all layers. As AI workloads become increasingly distributed, with inference and training tasks pushed to the edge, networks must provide not only bandwidth and reliability, but also real-time adaptability, support for in-network processing, and intelligent data handling. This evolution is particularly rele- vant across wireless and optical domains, where heterogeneous technologies, dynamic topologies, and latency-sensitive applications present new challenges and opportunities.
In this context, PLAIN – Programmable and Learning Architectures for Intelligent Networks aims to provide a focused venue for discussing how future network architectures must evolve to support AI- native operations. The workshop highlights the need for hardware/software co-design, where physical- layer programmability (e.g., P4, eBPF), AI algorithms, and orchestration strategies are jointly consid- ered to optimize the performance, flexibility, and efficiency of next-generation infrastructures.
While initiatives such as SIGCOMM’s NAIC explore networking for AI compute in cloud and data center environments, PLAIN focuses on the broader scope of AI across the edge-to-core continuum. It addresses scenarios where AI is embedded directly into the network, spanning adaptive protocols, telemetry-driven feedback loops, intent-based routing, and model distribution over programmable in- frastructures.
Emerging paradigms like Federated Learning demand networks that act as active participants in computation, dynamically adapting to workload placement, traffic patterns, and quality constraints. These applications highlight the need for a cross-layer approach, where the network is not just optimized for throughput or latency, but actively contributes to the execution of AI workflows.
A dedicated workshop on AI-native network architectures, co-located with ICIN 2026, is both timely and essential. It provides a platform for discussing how softwarized, programmable, and intelligent infrastructures can meet the demands of AI-enabled services. PLAIN promotes a cross-disciplinary vision, bridging systems, networks, and AI, to accelerate the development of scalable, sustainable, and intelligent next-generation networks.
Topics of interest (Not limited to)
Topics of interest include, but are not limited to:
- AI-driven design of network protocols and architectures
- Programmable and reconfigurable infrastructures for AI workloads
- Federated and edge AI enhanced by network awareness
- Orchestration of AI pipelines across heterogeneous domains
- Telemetry-driven network optimization and closed-loop control
- High-performance delivery and inference of AI models
- Experimental testbeds and platforms for AI-native networks
- Energy-efficient AI deployment at the edge and core
- AI-enabled automation in 6G RAN, core, and transport
- Digital twins for network monitoring, optimization, and control
- Resilience, security, and survivability of AI-native networks
- AI-driven orchestration of backhaul, midhaul, and fronthaul
- Open and disaggregated network architectures for AI integration
- In-network aggregation and distributed learning mechanisms
Contributors and Audience
This workshop is expected to attract contributors from both academia and industry, working at the intersection of AI, networking, and systems design. We welcome participants with expertise in the convergence of AI and transport networks, spanning optical and wireless domains, and from those advancing software-defined and programmable architectures.
Participants working on telemetry-based control loops, hardware/software co-design for AI accel- eration (e.g., using DPUs), or integration of AI mechanisms into the transport layer for RAN and backbone networks are especially encouraged to attend.
By targeting the enabling layers beneath AI, such as transport programmability, network telemetry, and orchestration, the workshop aims to bring together a community focused on rethinking the role of the network in supporting and accelerating AI pipelines.
This focus closely aligns with the themes of ICIN 2026, which emphasizes AI and sustainability in network and service management, and actively seeks contributions on topics such as SDN, orchestra- tion, telemetry, edge/federated learning, and green networking. As such, our workshop will offer ICIN attendees a forum to explore how transport-layer innovations can both enhance AI deployments and promote sustainable, NextG network architectures.
Tentative TPC
Here we report a list of potential program committee members who may also be interested to submit
a contribution to PLAIN 2026. Note that if the workshop is to be accepted, we commit to include a
wider and diverse TPC committee.
- Luca Valcarenghi, Full Professor at Scuola Superiore Sant’Anna
- Massimo Tornatore, Full Professor at Politecnico di Milano
- Walter Cerroni, Associate Professor at Universit`a di Bologna
- Andrea Sgambelluri, Researcher at Scuola Superiore Sant’Anna
- Alberto Gatto, Associate Professor at Politecnico di Milano
- Nicola Di Cicco, Operations Research Specialist at Optit S.r.l
- Alessandro Pacini, Telecom Cloud Engineer at BubbleRAN
- Davide Scano, R&D Engineer at Nokia
- Isabella Cerutti, Researcher at Joint Research Centre
- Jos ́e Santos, Researcher at Ghent University – imec
Format of the Workshop
The workshop will be held only in-person and structured as a half-day event, for approximately four
hours. The program will feature a keynote talk from a leading expert in the field, followed by up to
five peer-reviewed paper presentations.
To encourage interdisciplinary dialogue and engagement, we are also considering a panel discussion.
A coffee break will be included to facilitate informal networking among participants.
Table 1 illustrates a tentative agenda for the PLAIN workshop.
| Introduction | 10 mins |
| Keynote presentation | 1 hour |
| Technical session | 30-45 mins |
| Coffee break | 30 mins |
| Technical session | 30-45 mins |
| Panel discussion | 30-45 mins |
| Closing | 10 mins |
Publicity and Advertising Plan
We aim for the workshop to attract a significant number of high-quality submissions due to the high
relevance of the topics covered and the attention that such topics have gained in recent years in
academia and industry.
We also plan to use, among the others, the following ComSoc and IEEE mailing lists for our
advertising activities:
- netsoft-list@ee.ucl.ac.uk
- tccc-announce@comsoc.org
- tccn@comsoc.org
- cistc-distribution@comsoc.org
- commsoft@ieee.org
- csim@comsoc.org
- multicomm@comsoc.org
- ieeewtc@comsoc.org
- tcgcc@comsoc.org
- tc-comswtc@comsoc.org
- tchsn@comsoc.org
- cnom@comsoc.org
- etini@comsoc.org
- ontc@comsoc.org
We will also engage communities such as Open Networking Foundation (ONF), Optical Networking
Technical Committee (ONTC), and the SIGCOMM Community by posting the CFP on their channels
(e.g., the SIGCOMM Slack Channel).
Additionally, we will also capitalize on our personal LinkedIn profiles and communication channels
from our universities to promote the workshop.
Names and affiliations of the organization team
The Workshop Organizers are:
- Emilio Paolini – General Co-Chair, Scuola Superiore Sant’Anna, emilio.paolini@santannapisa.it
- Memedhe Ibrahimi – General Co-Chair,Politecnico di Milano, memedhe.ibrahimi@polimi.it
- Gianluca Davoli – TPC Chair, Universit`a di Bologna, gianluca.davoli@unibo.it
Short Biography of the Organizers
Bio: Dr. Emilio Paolini (Assistant Professor @ Scuola Superiore Sant’Anna) holds a PhD (with honors) in Emerging Digital Technologies at Scuola Superiore Sant’Anna, where he is currently an Assistant Professor. He is the recipient of NGI Enrichers Postdoctoral Fellowship, during which he worked on distributed AI infrastructure at Saint Louis University. He has been involved in several European Projects (BRAINE, CLEVER, SMARTY) as well as industrial funded projects (ComCast innovation fund, joint project with Saint Louis University). His current research is focused on evalu- ating the behavior and performance of AI models in large-scale distributed infrastructures and how to embed intelligence directly into the data plane, with a focus on the acceleration of AI models using the network infrastructure.
Bio: Dr. Memedhe Ibrahimi (Assistant Professor @ Politecnico di Milano) holds a PhD with honors in Information Technology from Politecnico di Milano. Dr. Ibrahimi has been involved in several industrial projects, as part of the BONSAI Lab at PoliMi, leading the development and im- plementation of Machine Learning – based solutions in communication networks. The main research interests include the application of machine learning in optical networks, and cross-layer network design considering capacity-scaling enablers such as Hollow-Core Fibers, Multi-Core Fibers, and Multi-band transmission.
Bio: Dr. Gianluca Davoli (Assistant Professor @ Universit`a di Bologna) received his PhD Degree in Electronics, Telecommunication, and Information Technology Engineering from the University of Bologna in 2021. His research work revolves around multiple aspects of softwarization in communica- tion infrastructures, including service orchestration over the Cloud continuum, with a focus on the Far Edge segment. He took part in the Organizing Committee of IEEE NFV-SDN 2023, IEEE NetSoft 2024 and 2025, and ONDM 2025. He was among the organizers of a workshop on Edge Network Softwarization (ENS) co-located with IEEE NetSoft since 2024.