Keywords in artificial intelligence
| Big data |
| Data of a very large size, typically to the extent that its manipulation and management present significant logistical challenges which makes it difficult to process it using on-hand data management tools or traditional data processing applications. A multitude of technology platforms may be involved in the generation or collection of this data, including network servers, digital imaging and communications in medicine (DICOM images) stored in picture archiving and communication system (PACS), electronic health records (EHRs), genomic data, personal computers, smart phones and mobile applications, and wearable devices and sensors. Artificial intelligence algorithms present the ability to search through such databases and extract the necessary data for specific task. |
| Algorithm |
| A set of rules provided to artificial intelligence that allows the machine to perform certain tasks, such as classification. |
| Machine learning |
| A field of artificial intelligence where a system is taught to interpret various features related to an objective and to flag it when it comes up. It is also trained to make predictions by recognising patterns. Automatic learning enables the machine to improve without continuous support and guidance by employing algorithms based on comparison logic, research and mathematical probability. |
| Classifier |
| A machine learning algorithm that sorts data into predefined categories. It can be: |
| Supervised which means it uses data that has been identified by the programmers to generate predictive models to identify unlabelled data. Often, the algorithm attempts to mimic the ability of a highly trained individual with domain expertise. Unsupervised which requires no data labelling. An output or a target is not defined. The computer looks for patterns or grouping in the data. This approach may be useful in facilitating analysis of big data. Reinforcement learning: the inputs and the outputs pairs are not specified, and the focus of the machine is to perform a task while learning from its own successes and mistakes to improve its performance. |
| Neural network |
| Neural networks mimic human brain function with each neuron performing its own simple calculation and the network formed by all of the neurons multiplying the potential of these calculations. This technology is particularly useful in predictive analysis, image recognition and speech processing. |
| Deep learning |
| A subcategory of machine learning where the system is made of digitised inputs, such as an image or speech, which go through multiple layers (ranging from 5 to 1000) of connected “neurons” (neural network), that progressively detect different features, and ultimately provides an output. The results from the first layer of neurons serve as a point of departure for calculating subsequent results. The connected neurons constitute the neural network with autodidactic quality. The system recognises patterns independently and makes predictions on a large quantity of information. |
| Natural language processing |
| A sub field of artificial intelligence dealing with the way to program computers to process and analyse large amounts of natural language data. Challenges are set by speech recognition, natural language understanding and natural language generation. It employs a sequence of mathematical processes and comparisons to identify the user's input, possibly correcting certain errors or applying synonyms, in order to identify all the information necessary to understand the needs of the user. |