Lune

ACL Findings2026

Infinite Babble: Inflating 3D Vision-Language Model Inference Overhead via Adversarial Geometric Perturbation

Shuoyang Sun, Jiaxin Hong, Yv Zhang, Kuofeng Gao, Hao Fang, Fan Mo, Bin Chen, Shu-Tao Xia

Abstract

3D Vision-Language Models (3D-VLMs) have emerged as the critical cognitive backbone for spatial intelligence, enabling precise reasoning over unstructured 3D data. While these models serve as the foundation for downstream robotics and embodied systems, their reliance on autoregressive decoding introduces a fundamental vulnerability regarding inference efficiency. In this work, we present Inflate3D, a novel adversarial framework designed to trigger computational and economic exhaustion in 3D-VLMs. Specifically, we exploit the model's sensitivity to untrusted 3D assets to hijack the generation process. Inflate3D operates by injecting imperceptible noise that forces the model into a state of pathological verbosity, effectively stalling the inference pipeline. Our approach comprises two synergistic strategies:

(1) a semantic-aware adversarial manipulation that leverages internal representations to selectively perturb semantically critical regions while preserving geometric structure, and (2) a trajectory disruption mechanism that manipulates token probabilities to suppress End-of-Sequence (EOS) emission, thereby prolonging decoding and inducing verbose outputs. Experiments on standard benchmarks show that In-flate3D amplifies output length and energy consumption by up to 6.45×, demonstrating a potent capability to drain system resources. These findings expose a critical blind spot in multimodal alignment, highlighting the urgent need to secure spatial foundation models against resource exhaustion attacks.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

Questions to start from

Lune can draw on

the full paper · related work

Get the answer

Free to start. No credit card required.

Lineage

Builds onCited byLeft to right by year; the more cited, the higher.

Builds on25

  1. An Image is Worth 16x16 Words: Transformers for Image Recognition at ScaleAlexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn et al.ICLR 2021 · 21,477 citations
  2. BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language ModelsJunnan Li, Dongxu Li, Silvio Savarese, Steven C. H. HoiICML 2023 · 7,873 citations
  3. 3D Gaussian Splatting for Real-Time Radiance Field RenderingBernhard Kerbl, Georgios Kopanas, Thomas Leimkühler, George DrettakisSIGGRAPH 2023 · 5,687 citations
  4. Gradient Surgery for Multi-Task LearningTianhe Yu, Saurabh Kumar, Abhishek Gupta, Sergey Levine et al.NeurIPS 2020 · 2,261 citations
  5. Overfitting in adversarially robust deep learningLeslie Rice, Eric Wong, J. Zico KolterICML 2020 · 935 citations

Related papers

  1. Inducing High Energy-Latency of Large Vision-Language Models with Verbose ImagesKuofeng Gao, Yang Bai, Jindong Gu, Shu-Tao Xia et al.ICLR 2024 · 79 citations
  2. Multimodal Resource-Exhaustion Attacks on Vision-Language Models via Joint Pixel-Prompt OptimizationZhaoxiong Ni, Yatie Xiao, Chi-Man Pun, Fei Peng et al.CCS 2026
  3. An Engorgio Prompt Makes Large Language Model Babble onJianshuo Dong, Ziyuan Zhang, Qingjie Zhang, Tianwei Zhang et al.ICLR 2025
  4. VLMInferSlow: Evaluating the Efficiency Robustness of Large Vision-Language Models as a ServiceXiasi Wang, Tianliang Yao, Simin Chen, Runqi Wang et al.ACL 2025 · 3 citations
  5. Crabs: Consuming Resource via Auto-generation for LLM-DoS Attack under Black-box SettingsYuanhe Zhang, Zhenhong Zhou, Wei Zhang, Xinyue Wang et al.ACL Findings 2025 · 6 citations
Dusk over the sea between two cliffs drawn in fine vertical lines