Glossary Terms

A generative adversarial network (GAN) is a deep-learning architecture that sets two neural networks against each other: a generator that produces candidate data and a discriminator that judges it. The competition between them pushes the generator toward increasingly convincing output. Sources: FDA Digital Health and AI Glossary; NIST AI Resource Center Glossary Read more

Generative artificial intelligence is the class of AI models that emulate the structure and characteristics of their input data in order to generate new, derived synthetic content. Source: FDA Digital Health and AI Glossary Read more

Geolocation is a high-level interface to location information associated only with the device hosting the implementation. Source: XRSI Standard XR Definitions (XR-001) Read more

Geospatial data is data about objects, events, or phenomena that have a location on the Earth’s surface. Source: XRSI Standard XR Definitions (XR-001) Read more

Ground truth is information established by direct observation rather than by inference. In AI development it is the trusted answer against which a model’s predictions are measured. Source: NIST AI Resource Center Glossary Read more

Haptics are technologies that deliver tactile feedback, such as vibration or resistance, so that a user can feel a digital interaction rather than only see it. In clinical simulation this is what allows a trainee to sense touching or palpating an organ or body part while practising a procedure. Sources: Unity Glossary; XRSI Standard XR… Read more

Head tracking detects the position and movement of a user’s head in real time so the view updates naturally as they look around. It provides both a stable virtual perspective and a simple method of input. Sources: Unity Glossary; XRSI Standard XR Definitions (XR-001) Read more

A head-mounted display (HMD), or headset, is a wearable device that places screens and optics directly in front of the eyes to deliver an immersive visual experience. Designs range from simple smartphone holders to high-end standalone units. Sources: Unity Glossary; XRSI Standard XR Definitions (XR-001) Read more

Hyperparameters are the settings used to configure a machine learning model, or to specify the algorithm that trains it, such as a learning rate. They are chosen before training rather than learned from the data. Source: NIST AI Resource Center Glossary Read more

Immersion is the psychological state of being deeply absorbed in a digital environment through multisensory engagement, to the point where awareness of the physical world recedes. It is the foundation on which a sense of presence is built. Sources: Unity Glossary; ACM Metaverse Glossary (Agrawal et al., 2020) Read more