SecureIoTM: Ultra-low-energy IoT Intrusion Detection Systems using Logic-based Tsetlin Machines

The SecureIoTM project seeks to transform the cybersecurity paradigm for interconnected devices. As devices, from smart fridges to trackers, join the IoT ecosystem, their security becomes vital. At its core is the Tsetlin Machine, a machine learning approach combining computational strength with energy efficiency. Distinct from conventional machine learning frameworks, the Tsetlin Machine offers a unique blend of computational vigor and energy conservation. With an exponential influx of devices, from critical medical equipment to everyday coffee makers, into the IoT framework, the SecureIoTM project is laser-focused on enhancing their resilience against cyber threats, thereby ensuring data integrity and privacy. This groundbreaking project is the result of collaboration among esteemed the Center of Artificial Intelligence Research at the University of Agder, Sørlandet Hospital Trust, and Newcastle University. The foundational bedrock of this initiative is based on four critical pillars of innovation. The first, Continual Learning, underscores a device's ability to persistently learn and adapt, retaining its prior knowledge and ensuring its ongoing evolution. The second, Adaptive Complexity, pertains to the Tsetlin Machine's capacity to adjust its model complexity based on the given task, balancing accuracy and energy efficiency. Cyber Resilience, the third, extends beyond threat detection, with the system proactively countering cyber threats in real-time, refining its defenses. The fourth, Seamless Integration, emphasizes the smooth integration of SecureIoTM's innovations into existing IoT structures, bolstering their inherent intelligence and security. The potential impact of the SecureIoTM initiative could revolutionize the cybersecurity realm. Should it achieve its ambitious goals, many of today's intrusion detection techniques for IoT might become antiquated, superseded by SecureIoTM's more adaptive, resilient, and energy-efficient methodologies.

The Internet of Things (IoT) has major security concerns because of its limited ability to dedicate system resources to security mechanisms. Combined with billions of resource-constrained devices in healthcare, agriculture, and industry, this offers cybercrime an economic incentive. IoT is heterogeneous systems, making traditional security considerably more difficult to implement. Intrusion Detection Systems (IDS) are particularly difficult since conventional stand-alone anomaly detection cannot run at low energy, required for power-hungry Wi-Fi.  We need IDS for IoT, as attackers can gain device access when passive security fails. Tsetlin Machines (TMs), a new machine learning (ML) technique based on the Tsetlin Automaton, can meet the energy requirements of IoT. TM hardware (HW) has proven over ten orders of magnitude less energy and faster learning compared to NNs, outperforming vanilla and binary forms of CNNs and Feed-Forward on well-established benchmarks. TMs extract common patterns in data via frequent pattern mining and resource allocation instead of output error minimization, which is prone to overfitting. With its ultra-low-energy architecture, TM is key to creating state-of-the-art IDS, but it requires study in three crucial areas. 1) Continual Learning for TM is crucial to online training as new sensor data or WIFI signals enter without catastrophic forgetting.  Machine learning, particularly TM, needs better techniques to prevent catastrophic forgetting. 2) Automatic tuning of hyperparameters for multimodal and temporal data is extremely difficult since IoT and its heterogeneity make individual devices modality-specific. This means that each device requires specific hyperparameter tuning, which is impossible for millions of IoT devices. 3) Cyber-Resilient Training is crucial to training robust models where attackers can't modify data undetected. An attack can readily deceive training without robust models, making anomaly detection ineffective.