Skip to main content

Table 1 Summary of the main definitions in the field of artificial intelligence (AI)

From: Artificial intelligence and hybrid imaging: the best match for personalized medicine in oncology

Term (alphabetical order) Definition
Computer-aided diagnosis and detection (CAD) systems A technology combining elements of artificial intelligence and computer vision that analyses imaging findings to estimate the likelihood that the feature represents a specific disease process (Suzuki 2012).
Clinical decision support system (CDSS) “A software designed to be a direct aid to clinical-decision making, in which the characteristics of an individual patient are matched to a computerized clinical knowledge base and patient-specific assessments or recommendations are then presented to the clinician for a decision” (Sim et al. 2001).
Deep learning (DL) A specific method of machine learning that incorporates artificial neural networks in multiple layers (i.e., “deep”) to iteratively learn from data, capable of autonomously identifying highly complex patterns in large datasets (Kirienko et al. in Press).
Distributed learning Health data network to share data that are made available to remote users by way of a query interface (Brown et al. 2010).
Intelligent system A systems that “learns from experience and makes appropriate choices” (Poole et al. 1998).
Machine learning A branch of artificial intelligence characterized by multiple hidden node layers that automatically learn data representations (i.e., “from experience”) by abstracting it in many ways (i.e., without being explicitly programmed) (Kirienko et al. in Press).
Natural language processing A subfield of artificial intelligence that classify and translate text, retrieve information, generate text, and interpret human language (Davenport & Kalakota 2019).
Neural networks (NN) “Also known as artificial NN, are networks using multiple layers of calculations to imitate the concept of how the human brain interprets and draws conclusions from information”(Handelman et al. 2018).
Holomics The gathering of genomic, radiomic, proteomic, clinical, immunohistochemical data, and their integration in predictive or prognostic models (Gatta et al. 2020).
Radiomics A mathematical method to extract from medical images handcrafted features (Sollini et al. 2019b).
Reinforcement learning “A computer program receives inputs from a dynamic environment in which it must achieve a certain goal (driving or playing against an adversary). During interaction with its problem space, the program receives feedback—rewards—which it attempts to maximize” (Kirienko et al. in Press).
Semi-supervised algorithm A machine learning method based on labeled and unlabeled input data used to learn a certain a task (van Engelen & Hoos 2020).
Statistical learning “A vast set of tools for understanding data which learn on the basis of some aspect of the statistical structure of elements of the input, primarily their frequency, variability, distribution, and co-occurrence probability” (James et al. 2013).
Supervised algorithms A process of an algorithm building a statistical model for predicting, or estimating, an output based on one or more inputs (training dataset) (James et al. 2013).
Texture A general term to describe the variation of the intensity (more in general the appearance) of a surface or a volume used to quantify regional descriptors in pattern recognition (e.g., cosmology, art, and medical imaging) (Sollini et al. 2018).
Unsupervised algorithms “A model the underlying structure or distribution in the data in order to learn more about the data” (James et al. 2013).