The environmental footprint of AI and reliance on cloud computing limit its sustainability and real-world applicability, particularly for use cases with strict privacy requirements (such as pharmaceuticals) or for applications lacking reliable connectivity (such as remote wildlife monitoring).

To address this, developing resource-aware & constrained AI is crucial. Our 180-minute session strategically merges the Danish Resource-Aware AI and Edge AI communities. Backed by the DIREC workstream on “CPS, IoT and Autonomous Systems” and P1 Program on “Green AI”, we aim to eliminate fragmentation, foster deep collaboration, and position Denmark as a global leader in sustainable, decentralized AI.

We assume a holistic, systems-level view of the AI landscape, recognizing that robust AI goes beyond model architectures to encompass data preparation, hardware-software co-design, and execution.

Programme

Lightning Research Talks (90 mins)
8 presentations (9 mins + 2 min Q&A), presenting either research, research ideas, or short tool presentations.

Lunch (60 mins)

Panel Discussion (60 mins)
A moderated debate on the future of Resource-Aware and Constrained AI.

Structured Networking (30 mins)
Participants will be split into targeted discussion groups based on their core interests (e.g., hardware, algorithms, societal impact) to foster cross-disciplinary networking and ideate future joint projects.

Speakers’ list

Luis Landa, Research Assistant, IT University of Copenhagen:
Solar-powered Computing
Presentation of Research Ideas

Josephine Plass-Wolter, PhD Student, University of Southern Denmark:
AI & Sustainability in Health
Presentation of Research Ideas

Purbak Sengupta, MSc Student, Aarhus University:
Predict Less, Know More: Event-First AI for Resource-Aware Wind Turbine Monitoring
Research Presentation

Mikkel Dahl, MSc Student / Research Assistant, University of Copenhagen:
CoDeQ+: Joint Quantization and Pruning for Extreme Structured Sparsity
Research Presentation

Alessandro Cerioli, PostDoc researcher, Technical University of Denmark:
Neural Network Optimization for Predictable Real-Time Edge Devices
Research Presentation

Roberto Morabito, Associate Professor, Technical University of Denmark:
[Title to be confirmed]
Presentation of Research Ideas

For the Panel Discussion:
Moderator: Jan Madsen, Head of Department, DTU Compute
Vivek Shah, Staff Engineer, Samsung
Dan Saattrup Smart, Alexandra Institute
Mahyar Tourchi Moghaddam, Associate Professor, SDU

Level

Intermediate: For attendees who have basic understanding or some experience with the subject but are not yet advanced.

Organizers

• Emil Njor (Assistant Professor, University of Southern Denmark) (Lead Organizer)
• Aparna Varde (Professor, University of Southern Denmark)
• Francesco Daghero (PostDoc, University of Southern Denmark)
• Sophia Wilson (PhD Student, University of Copenhagen)
  Raghavendra Selvan (Assistant Professor, University of Copenhagen)
• Akhil Arora (Assistant Professor, Aarhus University)
  Massimo Pavan (PostDoc, The Technical University of Denmark)
• Pınar Tözün (Associate Professor, IT University of Copenhagen)