Holistic Learning in Smart Environments with Liquid Spiking Neurosymbolic Networks

Wahab, Ahmad N. (2026) Holistic Learning in Smart Environments with Liquid Spiking Neurosymbolic Networks. Doctoral thesis, Birmingham City University.

Preview

Text Ahmad N. Wahab PhD Thesis_Final Version_Final Award June 2026.pdf - Accepted Version Download (13MB)

Abstract

Smart environments such as smart homes and smart cities produce sensor data that are dynamic, heterogeneous, sparse, and temporally irregular. Conventional machine learning models are generally developed for static datasets with fixed feature spaces and regular sampling patterns, limiting their suitability for environments in which sensors evolve, behaviours shift, and interdependencies emerge across space, time, and context. As a result, existing approaches often struggle to provide adaptive, scalable, and interpretable learning in real-world smart environments.

This thesis proposes a novel Liquid Spiking Neurosymbolic Network (LSNSN) framework for holistic learning in dynamic smart environments. The framework combines spiking neural networks, used as event-driven temporal attention mechanisms, with dynamic graph-based symbolic reasoning to model evolving relationships between sensors, events, and contextual entities. In doing so, it enables the integration of spatial, temporal, and semantic information within a unified neurosymbolic architecture capable of adaptive inference and structured interpretation.

The proposed framework is evaluated using synthetic, smart city, and smart home datasets, representing environments with differing temporal regularities, behavioural patterns, and levels of complexity. The results demonstrate that the framework can effectively model both cyclical and irregular event-driven data while supporting a fair degree of contextual reasoning over heterogeneous sensor streams. The research therefore contributes a novel adaptive and interpretable learning framework for smart environments, advancing the use of neurosymbolic and event-driven methods for real-time intelligent systems.

Item Type:	Thesis (Doctoral)
Dates:	Date Event 25 June 2026 Accepted
Uncontrolled Keywords:	Smart Environments, Holistic Learning, Spiking Neural Networks, Neurosymbolic
Subjects:	CAH11 - computing > CAH11-01 - computing > CAH11-01-01 - computer science CAH11 - computing > CAH11-01 - computing > CAH11-01-04 - software engineering CAH11 - computing > CAH11-01 - computing > CAH11-01-05 - artificial intelligence
Divisions:	Architecture, Built Environment, Computing and Engineering > Computer Science Doctoral Research College > Doctoral Theses Collection
Depositing User:	Louise Muldowney
Date Deposited:	06 Jul 2026 13:24
Last Modified:	06 Jul 2026 13:24
URI:	https://www.open-access.bcu.ac.uk/id/eprint/17101

Actions (login required)

View Item