AI@COLBY

Artificial Intelligence (AI) refers to the branch of computer science focused on making intelligent machines. There is no single, simple definition of AI due to its wide range of capabilities. AI systems can learn from data, reason, make decisions, and have a perception and a form of language comprehension. AI systems leverage a broad range of computational techniques, including machine learning, neural networks, natural language processing, and computer vision.

A Synergistic Ecosystem

The combination of big data, ICT advancements, high-speed computing, and neural networks has created a synergistic ecosystem that drives AI. For example:

• Big Data: Enables models to be trained with diverse and representative datasets.

• ICT Infrastructure and High-Speed Computing: Facilitates real-time data exchange and processing across networks and reduces the time required for training AI models, enabling rapid experimentation and development.

• Neural Networks: Enhance AI's ability to understand and replicate human-like intelligence.

Big Data

AI relies heavily on large volumes of data, often referred to as "big data," which serves as the foundation for training algorithms and improving their performance. The availability of vast, structured, and unstructured datasets, collected from sensors, websites, transactions, and social media, has significantly advanced AI. For instance, models like neural networks learn to recognize patterns and make predictions by analyzing this data.

Information and Communication Technologies (ICT)

Rapid advancements in ICT and the development of high-performance or advanced computing have been crucial in enabling modern AI. ICT provides the infrastructure for collecting, transmitting, and storing massive datasets, while fast computers equipped with powerful GPUs and TPUs accelerate the computationally intensive training of AI models. Cloud computing further democratizes access to AI tools, allowing organizations of all sizes to deploy AI applications.

Neural Networks

Serving as the backbone of AI, neural networks – inspired by the structure and functioning of the human brain – are a key component of AI. They consist of layers of interconnected nodes (neurons) that process and transmit information. Deep learning, a subset of machine learning, uses deep neural networks with many layers to solve complex problems such as image recognition, natural language processing, and autonomous driving. These networks excel at identifying intricate patterns and relationships in data, making them highly effective for a wide range of applications.

Generative artificial intelligence (GAI) is a type of artificial intelligence that “generates” new content (like text, images, or music) based on patterns learned from vast amounts of training data. It uses algorithms to generate outputs that resemble the original data but are new and unique, often mimicking human creativity.

A Large Language Model (LLM) is an advanced type of artificial intelligence based on complex statistical models originally designed for machine transcription and translation. LLMs are “trained” on vast amounts of data to enable them to predict a probable sequence of words in response to a prompt and produce coherent, contextually-relevant responses.

LLMs and generative AI are closely related because LLMs are a specific type of generative AI. Essentially, LLMs are the language-focused subset of generative AI, utilizing large datasets to generate relevant and contextually appropriate text responses.