Artificial Intelligence (AI) is everywhere in our modern world, and data driven AI models gain in popularity, yet again (Ukrainczyk 2004) also in engineering field of Construction and Building Materials. Artificial neural networks (ANNs) and machine learning methods have shown exceptional performance as predictive data driven computational tools, especially when used for pattern recognition, classification, and regression function estimation. They can capture complex highly non-linear interactions among many input/output parameters of a complex system without any prior knowledge about the nature of these interactions. The validity of a successfully trained AI model is determined by its ability to generalise its predictions beyond the training data and to perform well when it is presented with unfamiliar new data, typically from within the range of the input parameters used in the training.
Implement ANN and Machine learning (AI) models (e.g. in Matlab and/or Phyton), to be trained and validated for a range of data gathered from literature and/or lab measurements.
Following tasks should be focussed:
- Gathering data on a modelled (to be defined) process;
- Sorting the data into a table suitable for training of AI methods;
- Training, testing and validation of different AI methods, comparatively;
- Simulation research and sensitivity analysis
Possible topical focus:
Mix desing effects on prediction of various properties of building materials; such as: mechanical, rheological, durability, material characterisation etc.