PAISE 2025

Program

Workshop Program

This year, we received 14 high-quality submissions, and after extensive discussion, we accepted 6 Full and 2 Short papers. The following submissions have been accepted to PAISE 2025. Congratulations!

Date: 4th of June, 2025

Location: Room 3.0.2 in Building 3 on the Politecnico di Milano campus, located at Piazza Leonardo da Vinci 32, 20133 Milan.

Event Description Start End Duration
Introduction Introductory Remarks - PAISE Organizers 8:30 AM 8:40 AM 10 min
Keynote Keynote - Valeria Cardellini, Serverless without borders: Evolving across the Edge-Cloud continuum 8:40 AM 9:30 AM 50 min
Break - 1 9:30 AM 9:40 AM 10 min
Session 1 PAISE-01: SDFLMQ: A Semi-Decentralized Federated Learning Framework over MQTT. Amir Ali-Pour and Julien Gascon-Samson. 9:40 AM 10:05 AM 25 min
PAISE-02: Understanding the Performance and Power of LLM Inferencing on Edge Accelerators. Mayank Arya and Yogesh Simmhan. 10:05 AM 10:30 AM 25 min
Break - 2 10:30 AM 11:00 AM 30 min
Session 2 PAISE-03: Charon: An End-to-End Infrastructure for Connecting AI@Edge to HPC. Yongho Kim and Seongha Park and Swann Perarnau and Akhilesh Raj. 11:00 AM 11:15 AM 15 min
PAISE-04: Edge AI in the computing continuum: Consistency and Availability at Early Design Stages. Vincenzo Barbuto and Claudio Savaglio. 11:15 AM 11:40 AM 25 min
PAISE-05: Multi-Agent Reinforcement Learning for Workload Distribution in FaaS-Edge Computing Systems. Emanuele Petriglia and Federica Filippini and Michele Ciavotta and Marco Savi. 11:40 AM 12:05 PM 25 min
PAISE-06: Software Container-based Energy Estimation Models for ARM Architecture. Mohamed Anisse BELHADJ and Kods Trabelsi and Loïc Cudennec and Henri-Pierre Charles. 12:05 PM 12:30 PM 25 min
Lunch 12:30 PM 2:00 PM 90 min
Session 3 PAISE-07: SIMD Acceleration of Matrix-Vector Operations on RISC-V for Variable Precision Neural Networks. Gonzalo Salinas and Guilherme Sequeira and Alfonso Rodriguez and João Bispo and Nuno Paulino. 2:00 PM 2:25 PM 25 min
PAISE-08: Optimizing Speech Emotion Recognition with Dynamic Dilation Rates for Efficient Edge Deployment. Pin-Hsuan Lee and Jin-Shyan Lee. 2:25 PM 2:40 PM 15 min
Panel The Future of AI@Edge 2:40 PM 3:55 PM 75 min
Conclusion Closing Remarks, Feedback and Next Steps. 3:55 PM 4:15 PM 20 min
IPDPS Coffee Break 4:15 PM 4:30 PM 15 min

Keynote: Serverless without borders: Evolving across the Edge-Cloud continuum

Abstract: Serverless computing, particularly the Function-as-a-Service (FaaS) paradigm, has gained significant traction for its simplicity, scalability, and event-driven model. As the computing landscape shifts toward the heterogeneous Edge-Cloud continuum, extending serverless into this distributed environment presents both opportunities and challenges.

In this talk, we explore how the serverless model can support modern workloads, including AI-driven and latency-sensitive applications, across a continuum of devices and infrastructures. We address key challenges such as function scheduling and offloading, orchestration, and energy-efficient execution. We also outline emerging research directions and system-level considerations for building responsive, resource-aware, and sustainable serverless services.

As a case study, we present Serverledge, a decentralized, open-source FaaS platform built for the Edge-Cloud continuum and discuss how it addresses some of these challenges in practice, enabling borderless function execution across the compute continuum.

Valeria Cardellini

Valeria Cardellini is a professor of Computer Science and Engineering at the University of Rome Tor Vergata, Italy. She earned her PhD in Computer Science from the University of Rome Tor Vergata in 2001. Her research focuses on distributed software systems, including cloud and edge computing systems and services, their resource management and self-adaptation, and quality assurance. She has co-authored more than 100 publications in international journals and conferences. Additionally, she serves on the editorial boards of IEEE Transactions on Parallel and Distributed Systems and Elsevier Journal of Parallel and Distributed Computing, as well as on the technical program committees of system and performance conferences.





Panelists: The Future of AI@Edge

Prof. Biagio Cosenza Biagio Cosenza, University of Salerno

Prof. Biagio Cosenza is an associate professor at the University of Salerno (Italy). From 2015 to 2019, he was a senior researcher at TU Berlin (Germany), where he was the principal investigator of the DFG project Celerity and obtained his Habilitation. From 2011 to 2015, he was a postdoctoral researcher at the University of Innsbruck (Austria), where he contributed to the Insieme compiler and the DK-Plus scientific platform. He is a technical member of the Khronos Group industry standard body, a member of the Khronos SYCL Working Group, contributor to the UXL Foundation, and Senior Member of both ACM and the IEEE. His research has been funded by the project LibreRT (PRIN), LIGATE (EuroHPC JU), Celerity (DFG), and several industry-funded Ph.D. programs. His main research interests include programming models for heterogeneous systems, ranging from high-performance computing to embedded systems.

Prof. Nishant Saurabh Nishant Saurabh, Utrecht University

Prof. Nishant Saurabh is a tenured Assistant Professor in the Department of Information and Computing Sciences at Utrecht University, the Netherlands. He earned his Ph.D. in Computer Science from the University of Innsbruck in June 2021 and subsequently held a postdoctoral position at Klagenfurt University, Austria. His research focuses on hybrid distributed systems—spanning both classical and quantum computing—as well as cloud and edge computing, performance modeling, optimization, instrumentation, and observability. He has authored several papers in leading journals and conferences, including TPDS, JPDC, FGCS, IPDPS, and CCGrid, and contributes as a technical program committee member for various systems-related conferences and workshops. Additionally, he serves as an Associate Editor for Springer’s Journal of Cloud Computing and is a topic leader on Advanced Cloud Services for its editorial board. He is a member of IBM’s Quantum Computing for Material Science working group, contributing specifically to its HPC-Quantum integration and workload management subgroup. He also co-leads the CompSys Working Group on Digital Continuum, a Dutch national initiative advancing research in computer systems and networking.

Prof. Valeria Cardellini Valeria Cardellini, University of Rome Tor Vergata

Prof. Valeria Cardellini is a professor of Computer Science and Engineering at the University of Rome Tor Vergata, Italy. She earned her PhD in Computer Science from the University of Rome Tor Vergata in 2001. Her research focuses on distributed software systems, including cloud and edge computing systems and services, their resource management and self-adaptation, and quality assurance. She has co-authored more than 100 publications in international journals and conferences. Additionally, she serves on the editorial boards of IEEE Transactions on Parallel and Distributed Systems and Elsevier Journal of Parallel and Distributed Computing, as well as on the technical program committees of system and performance conferences.



Prof. Christophe Cérin Christophe Cérin, Sorbonne Paris Nord University

Prof. Christophe Cérin is a Computer Science Professor at Sorbonne Paris Nord University since 2005, focusing on high-performance computing, including grid, cloud, and edge computing. He develops middleware, algorithms, and tools for distributed system management and explores computing’s environmental impact. As an IEEE Member, he received the 2020 IEEE Technical Committee on Cloud Computing Research Innovation Award. He serves on the IEEE Transactions on Computers editorial board and as Associate Editor for Springer’s Journal of Cloud Computing. Starting September 2021, he joined Grenoble University for research with the Edge Intelligence program and INRIA Datamove team (Delegation INRIA programme). He recently chaired the FogSLA project (as local chair) on Fog orchestration and has industrial experience with the Wolphin, Wendelin, and Resilience projects related to cloud and big data.

Prof. Dragi Kimovski Dragi Kimovski, University of Klagenfurt

Prof. Dragi Kimovski is a tenured faculty at the Institute of Information Technology, University of Klagenfurt, Austria. He received his Ph.D. in Computer Science from the Technical University of Sofia in 2013 and completed his habilitation in distributed systems at the University of Klagenfurt. From 2009 to 2016, he was Assistant Professor at the University of Information Science and Technology in Ohrid, N. Macedonia, and subsequently held a research and teaching position at the University of Innsbruck from 2016 to 2018. Dr. Kimovski has published over 80 peer-reviewed articles in leading international journals and conferences. He has undertaken research stays at several renowned institutions, including the University of Michigan, University of Granada, University of Bologna, and Utrecht University. He has coordinated several national research projects and served as research coordinator and work package leader in multiple EU-funded projects, such as ENTICE, ASPIDE, and DataCloud. He is currently Chair of the SPEC Research Group on Edge Computing. His research focuses on cloud and edge computing, multi-objective optimization, and the application of decentralized machine learning in distributed environments. He regularly serves as editor and technical program committee member for top-tier international journals and conferences in the field of distributed and high-performance computing.