The integration of Artificial Intelligence (AI) with the Internet of Things (IoT), known as the Artificial Intelligence of Things (AIoT), enhances the devices’ processing and analysis capabilities and disrupts such sectors as healthcare, industry, and oil. However, AIoT’s complexity and scale are challenging for traditional machine learning (ML). Deep learning offers a solution but has limited testability, verifiability, and interpretability. In turn, theneuro-symbolic paradigmaddresses these challenges by combining the robustness of symbolic AI with the flexibility of DL, enabling AI systems to reason, make decisions, and generalize knowledge from large datasets better. This paper reviews state-of-the-art DL models for IoT, identifies their limitations, and explores how neuro-symbolic methods can overcome them. It also discusses key challenges and research opportunities in enhancing AIoT reliability with neuro-symbolic approaches, including hard-coded symbolic AI, multimodal sensor data, biased interpretability, trading-off interpretability, and performance, complexity in integrating neural networks and symbolic AI, and ethical and societal challenges.
Classification of clinical alarms is at the heart of prioritization, suppression, integration, postponement, and other methods of mitigating alarm fatigue. Since these methods directly affect clinical care, alarm classifiers, such as intelligent suppression systems, need to be evaluated in terms of their sensitivity and specificity, which is typically calculated on a labeled dataset of alarms. Unfortunately, the collection and particularly labeling of such datasets requires substantial effort and time, thus deterring hospitals from investigating mitigations of alarm fatigue. This article develops a lightweight method for evaluating alarm classifiers without perfect alarm labels. The method relies on probabilistic labels obtained from data programming—a labeling paradigm based on combining noisy and cheap-to-obtain labeling heuristics. Based on these labels, the method produces confidence bounds for the sensitivity/specificity values from a hypothetical evaluation with manual labeling. Our experiments on five alarm datasets collected at Children’s Hospital of Philadelphia show that the proposed method provides accurate bounds on the classifier’s sensitivity/specificity, appropriately reflecting the uncertainty from noisy labeling and limited sample sizes.
Run-time monitoring is a vital part of safety-critical systems. However, early-stage assurance of monitoring quality is currently limited: it relies either on complex models that might be inaccurate in unknown ways, or on data that would only be available once the system has been built. To address this issue, we propose a compositional framework for modeling and analysis of noisy monitoring systems. Our novel 3-value detector model uses probability spaces to represent atomic (non-composite) detectors, and it composes them into a temporal logic-based monitor. The error rates of these monitors are estimated by our analysis engine, which combines symbolic probability algebra, independence inference, and estimation from labeled detection data. Our evaluation on an autonomous underwater vehicle found that our framework produces accurate estimates of error rates while using only detector traces, without any monitor traces. Furthermore, when data is scarce, our approach shows higher accuracy than non-compositional data-driven estimates from monitor traces. Thus, this work enables accurate evaluation of logical monitors in early design stages before deploying them.