AI GLOSSARY - R
Definition: A versatile machine learning method capable of performing both regression and classification tasks. It utilises a multitude of decision trees at training time and outputs the class that is the mode of the classes (classification) or mean prediction (regression) of the individual trees.
Definition: In machine learning, ranking problems aim to predict an order among a list of items. It’s commonly used in recommendation systems where the task is to rank items according to their relevance to a user’s preferences.
Definition: A type of convolutional neural network that scans an image for potential objects by proposing regions (bounding boxes) and then runs a CNN on each of those regions in a second stage to classify the objects.
Definition: A class of neural networks where connections between nodes form a directed graph along a temporal sequence. This allows it to exhibit temporal dynamic behaviour and makes it suited to tasks like speech recognition or language modelling.
Definition: A type of machine learning technique that enables an agent to learn in an interactive environment by trial and error using feedback from its own actions and experiences.
Definition: A popular activation function used in neural networks, particularly in CNNs, defined as the positive part of its argument: 𝑓(𝑥)=𝑚𝑎𝑥(0,𝑥), where x is the input to a neuron.
Definition: A technique used to reduce the error by fitting a function appropriately on the given training set and avoid overfitting. This is done by adding a penalty term to the loss function.
Definition: A computer algorithm that mimics the learning process of humans and animals through rewards and punishments. It’s widely used in various applications, including robotics, gaming, and navigation.
Definition: A set of methods that allows a machine to be fed with raw data and to automatically discover the representations needed for detection or classification.
Definition: A type of convolutional neural network that uses skip connections or shortcuts to jump over some layers. Typical ResNets are implemented with double or triple layer skips that contain nonlinearities and batch normalisation in between.
Definition: The process of training a machine learning model or system that has already been trained with a new set of data to fine-tune the model or to improve its accuracy on new data.
Definition: A graphical plot used to show the diagnostic ability of a binary classifier system as its discrimination threshold is varied. It is commonly used to assess the performance of a classification model.
Definition: A standard way to measure the error of a model in predicting quantitative data. It is particularly useful when large errors are particularly undesirable.
Definition: A set of “if-then” rules used for making decisions or performing actions. Rule-based systems are used extensively in expert systems and other areas like medical diagnosis.
Definition: A variable whose possible values are numerical outcomes of a random phenomenon. In statistics and machine learning, random variables are used to model and manage stochastic or uncertain events.