Q1. (a) A company manufacturing rubber seals shown in Figure 1 would like to use Al for identifying good seals from three types of defects for quality control purpose. (0) (ii) Figure 1 Describe the type of data that could be of interest for training a suitable neural network and how this data could be acquired. [4 marks] Supposing there are 6 input features, design a suitable Neural Network architecture for identifying the defects using supervised learning method, and comment on the training data and the expected output. [4 marks] Supposing there are 6 input features, design a suitable Neural Network architecture for identifying the defects using unsupervised learning method, and comment on the training data and expected output. [4 marks] (iii)