4 papers
DySCo: Dynamic Sharding for Collaborative Edge-Cloud LLM Inference with Depth-Synchronized Batching
Jingpo Xu, Paul Joe Maliakel, Ivona Brandic +1
Pervasive intelligent applications are increasingly deployed on mobile and Internet of Things (IoT) edge devices. Consequently, Large Language Models (LLMs) are increasingly used t…
INAR-VL: Input-Aware Routing for Edge-Cloud Vision-Language Inference
Ahmed Šabanović, Paul Joe Maliakel, Ivona Brandić
Edge deployment of Vision-Language Models (VLMs) faces a tradeoff between latency and accuracy: cloud execution provides high-quality predictions but incurs communication delay and…
Characterizing LLM Inference Energy-Performance Tradeoffs across Workloads and GPU Scaling
Paul Joe Maliakel, Shashikant Ilager, Ivona Brandic
LLM inference exhibits substantial variability across queries and execution phases, yet inference configurations are often applied uniformly. We present a measurement-driven charac…
FLIGAN: Enhancing Federated Learning with Incomplete Data using GAN
Paul Joe Maliakel, Shashikant Ilager, Ivona Brandic
Federated Learning (FL) provides a privacy-preserving mechanism for distributed training of machine learning models on networked devices (e.g., mobile devices, IoT edge nodes). It…