Exploiting Efficiency Vulnerabilities in Dynamic Deep Learning Systems
This addresses security risks for real-world deployments of deep learning under strict latency and resource constraints, though it appears incremental as it builds on existing attack strategies and defense gaps.
This paper investigates security vulnerabilities in dynamic deep learning systems (DDLSs) that use input-adaptive computation for efficiency, revealing how adversaries can exploit these systems to cause excessive latency and energy usage, potentially leading to denial-of-service attacks.
The growing deployment of deep learning models in real-world environments has intensified the need for efficient inference under strict latency and resource constraints. To meet these demands, dynamic deep learning systems (DDLSs) have emerged, offering input-adaptive computation to optimize runtime efficiency. While these systems succeed in reducing cost, their dynamic nature introduces subtle and underexplored security risks. In particular, input-dependent execution pathways create opportunities for adversaries to degrade efficiency, resulting in excessive latency, energy usage, and potential denial-of-service in time-sensitive deployments. This work investigates the security implications of dynamic behaviors in DDLSs and reveals how current systems expose efficiency vulnerabilities exploitable by adversarial inputs. Through a survey of existing attack strategies, we identify gaps in the coverage of emerging model architectures and limitations in current defense mechanisms. Building on these insights, we propose to examine the feasibility of efficiency attacks on modern DDLSs and develop targeted defenses to preserve robustness under adversarial conditions.