This paper introduces Stochastic Intermittent Computing (STIC), a framework that integrates intermittent computing (ImC) and stochastic computing (SC) to enable always-on sensing in energy-harvested systems. STIC dynamically adjusts computational precision based on available energy, eliminating the need for non-volatile memory checkpointing traditionally used in ImC systems. By adapting precision in real-time, STIC ensures continuous operation even under severe power fluctuations, significantly improving energy efficiency and system resilience. Evaluation results demonstrate that STIC achieves substantial reductions in area, power, and energy consumption owing to the simplicity of SC and its tolerance to aggressive voltage scaling. Evaluations across multiple neural networks and charging traces confirm that STIC enables robust, low-power edge intelligence for resource-constrained environments.

RL-SEP is a reinforcement learning scheduler that optimizes neural network execution in energy-harvesting devices. By dynamically selecting quantization levels and early exit points, it improves active operation time by up to 11% over the reactive method while achieving 136% better accuracy-to-energy ratio and maintaining higher energy reserves. Testing on ResNet-18 and DenseNet-121 shows robust performance across various harvesting sources.