10 AI Terms You Must Know, Generative AI to Plugins

Artificial Intelligence (AI) continues to be developed by various companies. They even created AI products through various features, including chatbots. An AI chatbot from OpenAI that has the extraordinary ability to answer any question is called ChatGPT, currently the most popular AI-based tool that has been released in November 2022.

ChatGPT, this tool can help you with tasks such as writing poetry, resumes, and even planning holidays.

However, it took until the end of 2022 for most people outside the tech industry to start talking about AI. Keep in mind, AI chatbots are just one part of the AI ​​world. With AI being able to utilize ChatGPT to make work easier is a growing trend.

This is all because recent advances in machine learning are taking us to new heights. a major breakthrough with a major impact on almost every aspect of life. As people get used to the world of AI, of course new terms will appear everywhere.

Here are the AI ​​terms you need to know:

1. Machine learning (Machine learning / ML)

 If AI is the goal, then machine learning is how we get there. Machine learning is a field of AI computing, humans teach computer systems how to do things, training them to identify patterns and make predictions based on those patterns.

The data is run repeatedly through the algorithm, each time providing different input and feedback to help the machine learn and improve performance during training. Well, that's why large language models (LLM) use machine learning such as Bing Chat and ChatGPT.

2. Large language model (LLM)

Large language models (LLM) use machine learning techniques to help process language, so they can mimic the way humans communicate. LLM is a computer system inspired by the human brain, such as a collection of nodes and connections that simulate the neurons and synapses in the human brain. Models are trained using large amounts of text to learn patterns and relationships in language, to help the model use human words. Their problem-solving abilities can be used to translate languages, answer questions in chatbot form, summarize text, and even write stories, poetry, and computer code.

3. Generative AI

Generative AI harnesses the power of large language models to create new things, not just repeat or provide existing information. Not only that, generative AI learns patterns and structures, and then produces something similar but new. Generative AI can create things like images, music, text, video, and code. It can be used to create art, write stories, design products, and even help doctors with administrative tasks. However, it can also be used by bad actors to create fake news, or images that look like photos but are not real. This is because technology companies are developing ways to clearly identify AI-generated content.

4. Artificial intelligence (AI)

AI is a highly intelligent computer system, which can imitate humans in some ways. For example, understanding what people are saying, making decisions, translating language, analyzing whether something has a negative or positive tone, and even learning from experience. It is artificial in nature because its intelligence is created by humans using technology. Sometimes people say AI systems have digital brains, but AI is not a physical machine or robot. AI is a program that runs on a computer. The system works by simply feeding a very large collection of data through an algorithm which is a series of instructions to create a model.

5. Responsible AI

Responsible AI guides humans when trying to design systems that are safe and fair at every level, including machine learning models, software, user interfaces, and the rules and restrictions imposed on accessing applications. Responsible AI practices are an important element because AI systems are often tasked with helping make important decisions involving humans, such as in the fields of education and health. However, because AI is created by humans and trained using data from an imperfect world, it can reflect certain biases.

6. Hallucinations

Generative AI systems can create stories, poetry and songs, but sometimes humans want the results of generative AI to be based on truth. This is because the AI ​​system cannot differentiate between what is real and fake, generative AI can provide inaccurate responses. This phenomenon is what developers call hallucinations, or a more accurate term, fabrications. This is similar to when someone sees something that looks like the outline of a human face on the moon, and says that there really is a human on the moon. Developers are trying to resolve this issue through grounding, a technique of providing additional information from trusted sources to the AI ​​system, to improve the AI's accuracy on certain topics. Sometimes system predictions can also be wrong if the model does not have up-to-date information.

7. Multimodal Model

Multimodal Models can work with multiple data types or modes simultaneously. He can see pictures, hear sounds, and read words. In other words, multimodal models are true multitaskers! This model can combine all the information to perform tasks such as answering questions about images.

8. Prompts

Prompts are instructions entered into the system using language, images, or code to give tasks to the AI. Engineers and anyone who interacts with AI systems must design prompts carefully to get the desired results.

9. Copilots

Copilot is like a personal assistant that works with you in all kinds of digital applications, helping with tasks such as writing, coding, summarizing, and searching for information. Copilot can also help you make decisions and understand a lot of data. Recent LLM developments have made it possible for Copilots to understand everyday human language and provide answers, create content, or take actions, while humans work in different computer programs. Copilot was built with Responsible AI guidelines to ensure that the technology is safe, secure, and used for good.

10. Plugins

Plugins are similar to when you add an app to your smartphone: they exist to fill certain needs that may arise, allowing an AI app to do more without having to modify its underlying model. Plugins are what allow Copilot to interact with other software and services. Plugins can help AI systems access new information, perform complex mathematical calculations, or connect with other programs.

Share :