Multimodal AI foundation models for smart predictive and prescriptive maintenance

Smart predictive maintenance (SPdM) aims to predict when maintenance should be performed in an equipment, analysing its operational health status by monitoring its performance and condition during normal and faulty operations, to then predict which and when a failure might occur. 

Smart prescriptive maintenance (SPcM) is a step beyond SPdM, as it aims to not only predict which failure and when it will happen, but also to provide bespoke recommendations on what actions must be taken to avoid or mitigate losses. 

Furthermore, SPcM is a key enabler for Industry 5.0, aligning with its goals of people-centred, sustainability, and advanced technology integration, since it aims to provide actionable recommendations to optimise resources and enhance decision-making, helping industries to achieve higher efficiency, safety, resilience and productivity, while fostering collaboration between humans and machines, ultimately contributing to achieve UN Sustainable Development Goal 9 – “Industry, Innovation and Infrastructure”. 

They are achieved by employing artificial intelligence (AI) techniques to minimize unplanned downtime and maintenance costs, leveraging historical and multimodal data such as sensors, maintenance records, manufacturer’s manual etc., and combining advanced AI approaches from areas such as computer vision, natural language processing and time series processing. Thus, this project aims to develop advanced and state-of-the-art AI techniques leveraging the concepts of foundation models, data multimodality, physics-informed neural networks, trustworthy and ethical AI, to create a new class of SPcM tools that can be applied to different industry verticals with improved performance.

Open to any UK or international candidates.

You will need to meet the minimum entry requirements for our Vision, Speech and Signal Processing PhD programme.

We are looking for candidates that have the following skills: 

• Good knowledge in AI, especially for multimodal data, time series processing and foundation models.
• Previous experience in industry applications, especially in predictive/prescriptive maintenance. 
• Good programming in Python
• Expertise in data processing and analysis
• Good track record in AI development.

Institute for People-Centred AI