Faiyaz Doctor

Intelligent Connected Societies Research

In our increasing connected societies better insight into social processes such as labour mobility, socio-political sentiment, community care needs and strains on urban services can help to design better initiatives to manage citizens needs while mitigating the demands of an aging society, employment inequalities, urban traffic congestion and the impact of natural disasters. This insight is in part made possible through the increasing use of wireless sensing and distributed embedded computation which is motivating growth in connected computing artefacts enabling accessibility to diverse and emerging, multi-modal data sources. Examples of these data sources include geolocation, social media and online interaction from different personalized devices, mobile phone data, audio, visual, text, digitally sourced opinions, physiological signals, human emotion and sociometric sensor data. However, the use of Internet of Things (IoT) enabled artefacts by themselves lack intelligent processing capabilities such as pattern recognition, adaptivity, reasoning under uncertain data, search and recommendations. A key challenge here is to extract meaningful features and associations from these diverse data streams for modelling events, behaviors and conditions of interest.

My research interests lie in the design, simulation and application of smart ambient and human centred systems using nature inspired machine learning and Computational Intelligence (CI) techniques to facilitate interpretability, analysis and modelling of data for understanding human centered phenomena. Approaches such as agent-based modelling can be used for decomposing complex systems into their various actors and components to model their characteristics, and behavioirs. Generative and discriminative neural inspired algorithms can be used to reduce complex data dimensionality and extract useful features which can be used to model spatial-temporal patterns and predict the occurrence of phenomena. Real-world data sources are fraught with uncertainties pertaining to noise, human decisions, knowledge perception, reliability, trust and levels of agreement between different stakeholders. These sources of uncertainties can be managed and handled using fuzzy and probabilistic representation, aggregation and reasoning methodologies. Evolutionary algorithms can be applied to assess and optimize strategies through evolving multifaceted solutions against one or more objective criteria.

CI techniques can be applied to understand population mobility, financial and economic activities, social behavior, human sentiment, wellbeing, cybersecurity, political risk, education, welfare, geopolitics and environmental concerns.

PhD Topics

Related Publications:

My full list of publications on Goggle Scholar
My ResearchGate profile