Low-frequency antenna arrays enable obstacle-penetrating communications, but their large physical footprint—a 4 × 4 array at 150 MHz requires 16 m2—prevents practical deployment. DeformRF enables portable, deformable arrays that compress to backpack-size yet maintain beamforming performance when deployed on flexible substrates. Our key insight is that data-driven methods can predict complex electromagnetic behavior under deformation without real-time simulations. DeformRF combines: (1) a 260,000-sample synthetic dataset mapping deformations to EM characteristics, (2) physics-inspired ML models achieving >94% prediction accuracy on average, and (3) smartphone-based 3D reconstruction requiring no infrastructure. In real-world experiments, DeformRF maintains beamforming gains within 1 dB of optimal despite severe deformation, while baselines degrade by 4–10 dB. For emergency response scenarios, our 4 × 4 flexible array achieves ±5° direction-finding accuracy when tracking signals through buildings, enabling practical indoor deployment of large low-frequency arrays.

Resource-constrained IoT devices increasingly rely on deep learning models, however, these models experience significant accuracy drops due to domain shifts when encountering variations in lighting, weather, and seasonal conditions. While cloud-based retraining can address this issue, many IoT deployments operate with limited connectivity and energy constraints, making traditional fine-tuning approaches impractical. We explore this challenge through the lens of wildlife ecology, where camera traps must maintain accurate species classification across changing seasons, weather, and habitats without reliable connectivity. We introduce WildFiT, an autonomous in-situ adaptation framework that leverages the key insight that background scenes change more frequently than the visual characteristics of monitored species. WildFiT combines background-aware synthesis to generate training samples on-device with drift-aware fine-tuning that triggers model updates only when necessary to conserve resources. Our background-aware synthesis surpasses efficient baselines by 7.3% and diffusion models by 3.0% while being orders of magnitude faster, our drift-aware fine-tuning achieves Pareto optimality with 50% fewer updates and 1.5% higher accuracy, and the end-to-end system outperforms domain adaptation approaches by 20-35% while consuming only 11.2 Wh over 37 days—enabling battery-powered deployment. An open-source implementation of WildFiT is available at https://github.com/mmehdirk/WildFit.

While existing strategies to execute deep learning-based classification on low-power platforms assume the models are trained on all classes of interest, this paper posits that adopting context-awareness i.e. narrowing down a classification task to the current deployment context consisting of only recent inference queries can substantially enhance performance in resource-constrained environments. We propose a new paradigm, CACTUS, for scalable and efficient context-aware classification where a micro-classifier recognizes a small set of classes relevant to the current context and, when context change happens (e.g., a new class comes into the scene), rapidly switches to another suitable micro-classifier. CACTUS features several innovations, including optimizing the training cost of context-aware classifiers, enabling on-the-fly context-aware switching between classifiers, and balancing context switching costs and performance gains via simple yet effective switching policies. We show that CACTUS achieves significant benefits in accuracy, latency, and compute budget across a range of datasets and IoT platforms.