Large Language Models (LLMs) and Vision-Language Models (VLMs) remain highly vulnerable to adversarial attacks despite widespread adoption. Existing defenses typically require retraining, rely on heuristics, or fail under adaptive and out-of-distribution (OOD) conditions. We introduce EigenShield, a principled, inference-time, architecture-agnostic defense that leverages Random Matrix Theory (RMT) to suppress adversarial noise in high-dimensional embeddings. EigenShield uses spiked covariance modeling and a Robustness-based Nonconformity Score (RbNS) with quantile thresholding to isolate and preserve causal eigenvectors, filtering out adversarial components without model access or adversarial training. We develop a theoretical framework establishing conditions for asymptotic noise suppression and demonstrate effectiveness in both unimodal and multimodal settings. Empirically, EigenShield consistently improves robustness across threat models, reducing attack success rates (ASR) by up to 48% over state-of-the-art defenses, including adversarial training, UNIGUARD, CIDER, and input transformations. On jailbreak attacks, EigenShield lowers LLM ASR by up to 92.9% relative to undefended models. Under multimodal adversarial attacks, it reduces VLM ASR by up to 76.5%. Against adaptive attacks on LLMs, it achieves ASR reductions of up to 77.7%. In OOD settings, EigenShield maintains strong performance, reducing ASR by up to 88.4% for LLMs and 80.4% for VLMs.

Large Language Model (LLM) agents are now widely deployed in Ambient Intelligence (AmI) environments, where autonomous agents must sense, act, and coordinate at scale. As agent capabilities and interdependence increase, traditional reliability strategies such as isolated adaptive control, anomaly detection, or trust modeling have proven inadequate due to their fragmented and scenario-specific nature. Comprehensive architectures that enable integrated self-management, collective anomaly response, robust information dissemination, and privacy-preserving adaptation remain scarce. We propose a bio-autonomic framework for decentralized resilience in multi-agent LLM systems where a unified architecture systematically applies principles from biological autonomic systems to LLM-based multi-agent environments. Specifically, each agent implements an autonomic control loop, formally structured as Monitor-Analyze-Plan-Execute over a shared Knowledge base (MAPE-K), for self-regulation. At the system level, the framework integrates immune-inspired anomaly detection using the Dendritic Cell Algorithm, probabilistic computational trust, decentralized gossip for robust information sharing, and federated learning with homomorphic encryption for collaborative, privacy-preserving adaptation. This holistic approach enables LLM agent ecosystems to self-organize, detect and isolate faults, and collectively adapt as system complexity increases. Empirical evaluations show that our framework achieves substantially improved resilience and recovery compared to state-of-the-art multi-agent baselines.

Autonomous edge computing in robotics, smart cities, and autonomous vehicles relies on the seamless integration of sensing, processing, and actuation for real-time decision-making in dynamic environments. At its core is the sensing-to-action loop, which iteratively aligns sensor inputs with computational models to drive adaptive control strategies. These loops can adapt to hyper-local conditions, enhancing resource efficiency and responsiveness, but also face challenges such as resource constraints, synchronization delays in multimodal data fusion, and the risk of cascading errors in feedback loops. This article explores how proactive, context-aware sensing-to-action and action-to-sensing adaptations can enhance efficiency by dynamically adjusting sensing and computation based on task demands, such as sensing a very limited part of the environment and predicting the rest. By guiding sensing through control actions, action-to-sensing pathways can improve task relevance and resource use, but they also require robust monitoring to prevent cascading errors and maintain reliability. Multi-agent sensing-action loops further extend these capabilities through coordinated sensing and actions across distributed agents, optimizing resource use via collaboration. Additionally, neuromorphic computing, inspired by biological systems, provides an efficient framework for spike-based, event-driven processing that conserves energy, reduces latency, and supports hierarchical control-making it ideal for multi-agent optimization. This article highlights the importance of end-to-end co-design strategies that align algorithmic models with hardware and environmental dynamics, improve cross-layer inter-dependencies to improve throughput, precision, and adaptability for energy-efficient edge autonomy in complex environments.

We propose a disruptively frugal generative sensing approach for LiDAR that generates, rather than senses, parts of the environment that are either predictable based on extensive training or have limited impact on overall prediction accuracy. Our generative pre-training strategy for this purpose, radially masked autoencoding (R-MAE), also allows focusing on radial segments of the data, which captures spatial relationships and distances between objects more effectively than conventional procedures. As a result, the proposed methodology not only reduces sensing energy but also improves prediction accuracy. Our evaluations on Waymo and nuScenes show that our approach achieves over a 5% average precision improvement in detection tasks across datasets and over a 4% accuracy improvement when transferring domains from Waymo and nuScenes to KITTI. Our method achieves up to 3.17% and 2.31% improvements in mean average precision (mAP) and NDS, respectively, in nuScenes. Even with 90% radial masking, it surpasses baseline models by up to 5.59% in mAP and mean average precision with heading (mAPH) across all object classes in the Waymo dataset. Resultantly, R-MAE reduces the sensing energy by ∼9.1× than the conventional LiDAR processing despite computational overheads. Codes are available: https://github.com/sinatayebati/Radial_MAE.

We propose SPARC, a lightweight continual learning framework for large language models (LLMs) that enables efficient task adaptation through prompt tuning in a lower-dimensional space. By leveraging principal component analysis (PCA), we identify a compact subspace of the training data. Optimizing prompts in this lower-dimensional space enhances training efficiency, as it focuses updates on the most relevant features while reducing computational overhead. Furthermore, since the model’s internal structure remains unaltered, the extensive knowledge gained from pretraining is fully preserved, ensuring that previously learned information is not compromised during adaptation. Our method achieves high knowledge retention in both task-incremental and domain-incremental continual learning setups while fine-tuning only 0.04% of the model’s parameters. Additionally, by integrating LoRA, we enhance adaptability to computational constraints, allowing for a tradeoff between accuracy and training cost. Experiments on the SuperGLUE benchmark demonstrate that our PCA-based prompt tuning combined with LoRA maintains full knowledge retention while improving accuracy, utilizing only 1% of the model’s parameters. These results establish our approach as a scalable and resource-efficient solution for continual learning in LLMs.

Most scientific discoveries depend on identifying novel signals hidden in massive, noisy datasets generated by modern experiments. Traditional novelty detection methods are often insufficient in speed, robustness, and adaptability to resource-constrained environments. We discuss a perspective on a hierarchical framework for multi-level novelty detection spanning sensor signals, feature representations, and model outputs. At the signal level, we discuss analog circuits that extract statistical densities and moments in real-time, enabling interpretable and energy-efficient filtering. At the feature level, we introduce Likelihood Regret, an unsupervised measure that detects anomalies by retraining generative models on shared latent representations, with optimizations for embedded deployment. At the output level, we leverage predictive uncertainty, applying both compute-efficient Monte Carlo reuse and Monte Carlo-free techniques like evidential learning and conformal inference. Our framework demonstrates how integrating novelty detection across the sensing-to-inference can accelerate insights in domains such as high-energy physics.

Complex sensors such as LiDAR, RADAR, and event cameras have proliferated in autonomous robotics to enhance perception and understanding of the environment. Meanwhile, these sensors are also vulnerable to diverse failure mechanisms that can intricately interact with their operational environment. In parallel, the limited availability of training data on complex sensors affects the reliability of their deep learning-based prediction flow, when their prediction models fail to generalize to environments not adequately captured in the training set. To address these reliability concerns, this paper introduces STARNet, a Sensor Trustworthiness and Anomaly Recognition Network designed to detect untrustworthy sensor streams that may arise from sensor malfunctions and/or challenging environments. STARNet employs the concept of likelihood regret (LR) for continuous evaluation of the trustworthiness of sensor streams. We tailor the framework to resource-constrained edge devices with two settings: a gradient-free framework suit- able for low-complexity hardware with fixed-point precision capabilities, and a low-rank tunability of underlying models for LR extraction, reducing the extraction workload. Through extensive simulations, we demonstrate the efficacy of STARNet in detecting untrustworthy sensor streams in unimodal and multimodal settings. In particular, the network shows superior performance in addressing internal sensor failures, such as cross-sensor interference and crosstalk. In diverse test scenarios involving adverse weather and sensor malfunctions, we show that STARNet enhances prediction accuracy by approximately 15% by filtering out untrustworthy sensor streams. STARNet is publicly available at https://github.com/nstrndrbi/STARNet.