CVAIApr 25, 2024

Energy-Latency Manipulation of Multi-modal Large Language Models via Verbose Samples

arXiv:2404.16557v118 citationsh-index: 17
Originality Incremental advance
AI Analysis

This exposes a security risk for MLLM deployment, potentially harming service availability through resource exhaustion, but it is incremental as it builds on known adversarial attack concepts.

The paper tackles the vulnerability of multi-modal large language models (MLLMs) to energy-latency attacks by crafting imperceptible perturbations called verbose samples, which extend generated sequence lengths by up to 50% in experiments, increasing computational costs.

Despite the exceptional performance of multi-modal large language models (MLLMs), their deployment requires substantial computational resources. Once malicious users induce high energy consumption and latency time (energy-latency cost), it will exhaust computational resources and harm availability of service. In this paper, we investigate this vulnerability for MLLMs, particularly image-based and video-based ones, and aim to induce high energy-latency cost during inference by crafting an imperceptible perturbation. We find that high energy-latency cost can be manipulated by maximizing the length of generated sequences, which motivates us to propose verbose samples, including verbose images and videos. Concretely, two modality non-specific losses are proposed, including a loss to delay end-of-sequence (EOS) token and an uncertainty loss to increase the uncertainty over each generated token. In addition, improving diversity is important to encourage longer responses by increasing the complexity, which inspires the following modality specific loss. For verbose images, a token diversity loss is proposed to promote diverse hidden states. For verbose videos, a frame feature diversity loss is proposed to increase the feature diversity among frames. To balance these losses, we propose a temporal weight adjustment algorithm. Experiments demonstrate that our verbose samples can largely extend the length of generated sequences.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes