1072 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
Russell (BBC–Reith Lectures, 2021) the last couple of
years has a been a watershed moment for the application of
AI. Combined with low–powered computing and enabled
by faster network bandwidths to access cloud–based solu-
tions, AI–enabled tools and platforms are now accessible to
anyone with a smartphone running advanced algorithms
and computations. Voice–assisted devices such as Apple’s
Siri, Amazon’s Alexa and Android’s Google assistants can
understand voice commands and undertake tasks seam-
lessly with relative accuracy.
AI systems can be classified into Narrow and General
(Figure 2). Narrow AI is widely used in today’s systems
and tools. Narrow AI is trained on familiar data available
in existing datasets (internet, books, images etc..). General
AI aims to behave like humans, solving unfamiliar prob-
lems and datasets. The LLMs use generative AI, a type of
machine learning that can be creative and generate new
outcomes. Generative AI is a step towards developing
General–Purpose AI.
The Application
Advanced algorithm techniques, such as deep neural and
machine learning, use different levels of software architec-
ture that leverage data to learn and improve predictability.
These base software languages are core to building an AI
platform. Figure 3 resents a schematic of the levels of soft-
ware architecture required to build an AI platform. Data is
critical to all three levels of software architecture, especially
the access to good training data to teach the software to
identify, solve, and provide solutions to various tasks.
• The core is deep learning, which comprises neural
networks. The “Deep” in deep learning refers to a
neural network comprised of more than three lay-
ers—including the inputs and the output—and
Figure 1. Moore’s law—The number of transistors on a microchip (Moore’s Law, 2022)
Russell (BBC–Reith Lectures, 2021) the last couple of
years has a been a watershed moment for the application of
AI. Combined with low–powered computing and enabled
by faster network bandwidths to access cloud–based solu-
tions, AI–enabled tools and platforms are now accessible to
anyone with a smartphone running advanced algorithms
and computations. Voice–assisted devices such as Apple’s
Siri, Amazon’s Alexa and Android’s Google assistants can
understand voice commands and undertake tasks seam-
lessly with relative accuracy.
AI systems can be classified into Narrow and General
(Figure 2). Narrow AI is widely used in today’s systems
and tools. Narrow AI is trained on familiar data available
in existing datasets (internet, books, images etc..). General
AI aims to behave like humans, solving unfamiliar prob-
lems and datasets. The LLMs use generative AI, a type of
machine learning that can be creative and generate new
outcomes. Generative AI is a step towards developing
General–Purpose AI.
The Application
Advanced algorithm techniques, such as deep neural and
machine learning, use different levels of software architec-
ture that leverage data to learn and improve predictability.
These base software languages are core to building an AI
platform. Figure 3 resents a schematic of the levels of soft-
ware architecture required to build an AI platform. Data is
critical to all three levels of software architecture, especially
the access to good training data to teach the software to
identify, solve, and provide solutions to various tasks.
• The core is deep learning, which comprises neural
networks. The “Deep” in deep learning refers to a
neural network comprised of more than three lay-
ers—including the inputs and the output—and
Figure 1. Moore’s law—The number of transistors on a microchip (Moore’s Law, 2022)
