Definitions of artificial intelligence (AI) and related technologies such as Large Language Models (LLMs) vary across news and social media channels. You’re not alone if it’s difficult to understand and connect the dots. There are several distinct technologies within the AI bucket, so let’s take a simplified look.
Artificial intelligence (AI) is a branch of computer science that creates and trains systems or models to imitate how humans perform intellectual tasks. Models are mathematical tools that help computers become more capable. When trained with accurate, high-quality data, AI models can enable computers to:
-
Recognize images such as food, furniture, cars, faces and expressions, buildings, landmarks, text, handwriting, medical images, artwork, hand gestures, and more.
-
Make decisions to sort and categorize data, determine the speeds and directions of autonomous vehicles, detect fraud, suggest options for medical treatment, and translate text from one language to another.
-
Solve problems such as differentiating between real and junk emails, summarizing key themes across large volumes of text, determining positive or negative sentiments throughout news and social media content, optimizing global supply chains to transport consumer goods, or recommending films to stream.
-
Communicate in a conversational way to answer customer questions in chat sessions or recognize what someone is saying.
FactSet has been using AI throughout its solutions since 2007.
Subsets of Artificial Intelligence
Generative AI models generate new content, including text, images, and sounds. ChatGPT is an example of generative AI for text. Computer engineers use large datasets to train the models to learn patterns and structures. The models then use their learned knowledge to combine or modify the data (e.g., text, images, music) and produce new, original content. For example, after the completion of training, generative AI can create images for video games, talk like a real person for phone support, or develop melodies and sound effects for movies.
Large Language Models (LLMs) process and generate human text at an advanced level. They are trained on enormous amounts of text data to learn the statistical patterns, grammar, syntax, and the meaning of human language. LLMs have a remarkable capability to understand and generate text that is coherent and contextually relevant. LLMs can also translate text to different languages, summarize articles, answer questions from documents, and perform a variety of other language-related tasks.
FactSet has been using LLMs such as Bloom, Google’s BERT, and T5 across its suite of products and services since 2018 and is now enhancing its services with the newest and most advanced LLMs like ChatGPT and GPT4.
Machine Learning (ML) enables computers to learn from data without the involvement of a human programmer throughout the process. It involves teaching a computer both to recognize patterns in data and to make decisions based on those patterns. Algorithms (i.e., step-by-step instructions a computer follows to perform a task or solve a problem) are part of the training process. The computer can improve its performance as it is fed more data and can refine its ability to recognize patterns and adjust its decisions accordingly. In simple terms, Machine Learning helps computers to learn from experience and make decisions based on that learning.
Natural Language Processing (NLP) is a field of study that helps individuals communicate with computers in a natural, intuitive way instead of through the complex programming languages that computer engineers use. It enables computers to “understand” the meaning of words and sentences, recognize patterns in language, and respond in a human-like manner (writing or speaking). Its applications range from chatbots and virtual assistants to language translation and sentiment analysis. The bottom line: NLP allows machines to understand and communicate with humans using conversational language.
Interesting in Learning More?
Visit FactSet Artificial Intelligence and view additional Insight blog articles about AI:
AI and the Investment Management Process (factset.com)
AI Quick Bite: Embeddings and Large Language Models (factset.com)
Reducing Hallucinations from Generative AI (factset.com)
ChatGPT Facts vs. Hype: What Are Large Language Models? (factset.com)
Understanding ChatGPT: The Promise and Nuances of Large Language Models (factset.com)
This blog post is for informational purposes only. The information contained in this blog post is not legal, tax, or investment advice. FactSet does not endorse or recommend any investments and assumes no liability for any consequence relating directly or indirectly to any action or inaction taken based on the information contained in this article.
Post Comment
