SuRE 2026
Sustainability and Resource-Efficiency of Artificial Intelligence 2026

Table of Contents

• Preface
  Summary: There were 27 papers submitted for peer-review to this workshop. Out of these, 8 papers were accepted for this volume.

Paper Session I: Power Usage

• Safety-Constrained Contextual Bandit for Dynamic Power Management
  Yasin Esfandiari , Karan Rajshekar , Cicy Agnes , Hannah Stein , Sabine Janzen , Wolfgang Maass
• WattGPU: Predicting Inference Power and Latency on Unseen GPUs and LLMs
  Mauricio Fadel Argerich , Jonathan Fürst , Marta Patiño-Martínez
• WattLayer: Get Layers Right to Estimate Inference Energy of Neural Networks
  Adrien Sardi , Marie Line Alberi-Morel , Sara Alouf , Frédéric Giroire , Joanna Moulierac

Paper Session II: Efficient Models & Deployment

• CompressKV: Semantic-Retrieval-Guided KV-Cache Compression for Resource-Efficient Long-Context LLM Inference
  Xiaolin Lin , Jingcun Wang , Olga Kondrateva , Yiyu Shi , Bing Li , Grace Li Zhang
• TinyHybrid: Ultra-Efficient CNN-Transformer Architecture for Edge-Deployable Brain Tumor Classification
  Krystian Krukowski , Pawel Powroznik , Maria Skublewska-Paszkowska
• ENAS: An Efficient Hardware-Aware Neural Architecture Search Framework for TinyML on Resource-Constrained Microcontrollers
  Mohd Moin Khan , Naman Srivastava , Pandarasamy Arjunan

Short Paper Session

• Accounting for Bias Enables Sustainable LLM Evaluation
  Harshita Katoch , David Antony Selby , Gerrit Großmann , Sebastian Vollmer
• Efficient Vision Models for Jetson: Steel Classification via Knowledge Distillation
  Yasin Esfandiari , Karan Rajshekar , Sabine Janzen , Wolfgang Maass