New Method Improves Query-Efficient Adversarial Attacks on AI Systems
Researchers have proposed a new adversarial attack method called Latent Geometric Chords (LGC) that targets AI decision boundaries using geometric search within compressed semantic spaces. The method introduces a Residual-based Adversarial Generation (RAG) mechanism to improve visual fidelity and expand the search space, addressing limitations of prior pixel-wise and latent-space attack approaches. The work is significant because it demonstrates that even adversarially trained robust models can be compromised with minimal perceptible image distortion and relatively few queries.
A preprint submitted to IEEE Transactions on Information Forensics and Security introduces Latent Geometric Chords (LGC), a decision-based black-box adversarial attack framework designed to overcome key weaknesses in existing methods. Prior pixel-wise attacks tend to produce visually unnatural artifacts, while latent-space approaches are constrained by low-dimensional manifolds and reconstruction errors. LGC addresses these issues by performing a curvature-aware geometric search within a compressed semantic manifold, with a companion variant called LGC-H also proposed. The central innovation, the RAG mechanism, isolates semantic perturbations as geometric chords and overlays them directly onto source images, effectively doubling the usable search space dimensions while preserving image quality. Experimental results report an SSIM above 0.99 and LPIPS below 0.01 at 5,000 queries, indicating near-imperceptible perturbations. The method demonstrates strong cross-dataset transferability and outperforms state-of-the-art baselines, including successfully attacking adversarially hardened models. Source code has been made publicly available by the authors.
What's missing
The paper is a preprint and has not yet completed peer review at IEEE Transactions on Information Forensics and Security. Key limitations not detailed in the abstract include the specific model architectures and datasets tested, computational costs relative to baselines, and whether the attack generalizes across diverse real-world deployment conditions. The threat model's practical applicability — including how realistic the assumed query access is in deployed systems — is not addressed in the available abstract.
What different sources said
- arXiv cs.LGCenter
Latent Geometric Chords for Query-Efficient Decision-Based Adversarial Attacks
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.