AI Technology Types
AI Technology Types are technology elements used to create the use cases.
The AI technology types listed here reflect the technology types emerging in the maritime industry right now. However, we anticipate that this list will grow over time as the adoption of AI increases.
If you are looking for an AI technology type that is not currently listed here, please get in touch.
The computing infrastructure allows different elements of the AI to perform on-board (edge computing) and onshore (private/public cloud) analysis and decision-making.
When used for asset health management, digital twins consist of AI that can perform anomaly detection, fault detection and isolation, diagnostics, prognostics to generate servicing, maintenance actions including operational advisories for the asset.
A category of AI that focuses on abstracting/learning multiple examples of data (i.e. multivariate data) relevant to the application domain to perform analysis, generate insights and decision making
This is a category of AI that combines the knowledge-based and data-driven types of AI to maximise the accuracy, generalisation capability, consistency of the analysis, decision making and insight generation properties of the AI system
This is a category of AI where expert knowledge (e.g. engineering principles, physical understanding of phenomena, mathematical functions, etc.) is translated to a knowledge base for use by an inference reasoner for analysis and decision making.
This is a sub-category of data-driven AI, whose main objective is the analysis and synthesis of natural language in the form of text and speech. NLPs are normally coupled with other types of AI to perform analysis and decision making (e.g. NLP can read and analysing system alarm codes and logs for further processing by a fault detection type of AI)
This is a sub-category of data driven AI, whose functionality tries to mimic the human brain by having multiple neurons arranged via input layer, hidden (computation) layers and output layer to perform analysis and decision making. There are multiple variants of neural networks such as convolutional neural networks, deep neural networks, etc.
Most AI applications perform sensor fusion which is the process of combining disparate, varying data types (e.g. a digital twin can use pressure, temperature, vibration, alarm logs. An autonomous navigation system will use radars, cameras, GNSS, engine speed) to perform the analysis and decision making relevant to the application