Communication & Perceiving (NLP, CV)

Certified AI Developer (CAID) · 5 questions

• Your team is building a computer vision system to identify defects in manufacturing parts on a conveyor belt. A senior engineer recommends using a traditional machine learning approach like Support Vector Machines (SVMs) with hand-crafted descriptors like HOG (Histogram of Oriented Gradients). You advocate for a Deep Learning approach using a Convolutional Neural Network (CNN). What is the primary advantage of your proposed deep learning approach over the traditional pipeline?
• You are designing a Convolutional Neural Network (CNN) for facial recognition. As the feature maps pass deeper into the network, the computational load threatens to overwhelm your hardware. Which type of layer should you insert to downsample the feature maps, reducing their width and height while retaining the most important visual information?
• Computers are great at math, but they don't understand human words out of the box. In modern Natural Language Processing (NLP), how do we represent words so that models can process them mathematically while capturing their semantic meanings and relationships?
• A clinic wants to implement an AI system that analyzes chest X-ray scans to automatically flag potential cases of pneumonia. Given that the input data consists of 2D pixel grids with local spatial relationships, which neural network architecture is best suited for this task?
• In computer vision architectures, what is the primary role of a convolutional layer within a Convolutional Neural Network (CNN)?