Week 12 : Neural Networks 1 Neural Networks
The basic building blocks of the neural networks, their construction mechanisms and methodology for learning are discussed.
Abstract
Neural Networks are an important class of models in machine learning and computer vision. Neural networks have produced, and continue to produce, impressive results in the broad spectrum of vision problems (e.g., classification, object detection, segmentation, and many many others). As with any supervised learning, neural networks allow one to define a certain parametric model and then learn (optimize) parameters of that model based on data in the form of input-output pairs (e.g., images and labels). In order to do this one typically needs to define a problem dependent objective or a loss function which measures the severity of miss-prediction for each example in the training dataset. In what follows, we will discuss the basic building blocks of the neural networks, their construction mechanisms and methodology for learning.