Artificial Intelligence (AI) Jargon Demystified

If you are curious about the difference between the chatbots and AI-assistants, or what is meant by deep learning vs. machine learning vs. neural networks, you have come to the right place.? Today we are going to discuss some common AI related terms so that next time when you are in an AI related conversation you don't feel left out, here we go.

Chatbots and AI-Assistant:? The terms "chatbot" and "AI-assistant" are often used interchangeably, but they refer to different type of application and capabilities. Chatbots are software applications which are designed to simulate conversation through text or speech based on user inputs.? Chatbots tend to use rule-based systems or simple AI to respond to the query.? AI assistants also known as virtual assistants or intelligent personal assistants are software applications that use AI to perform tasks and services to improve efficiency on behalf of users. AI-assistants continuously learn from user interactions to improve responses, progressively using more complicated algorithms.

In most AI discussions, terms like machine learning, deep learning and neural networks are almost always brought up.? They are related to each other, but have distinct meanings.

• Machine learning: focuses on developing algorithms that learn from making decisions based on data.? There are different approaches like supervised, unsupervised, and reinforcement learning which we will discuss in our next blog.
• Deep learning: focuses on a subset of machine learning that uses neural networks to automatically learn from a large amount of data.? Deep learning excels in complex tasks like image and speech recognition.
• Neural Networks: Is a method of data processing for machines in a way inspired by the human brain’s function.

Now, let's discuss some other AI jargon:

Algorithms:? Sets of instructions that a program follows in order to give you desired results.

Bias:? In an AI context, bias refers to incorrect results produced by the algorithm because of insufficient, incomplete, and/or generally flawed data.

Data Mining:? The process of combing through large sets of data to find patterns or trends.

Large Language Model (LLM): a LLM is a deep learning algorithm that is trained on a massive data set to generate and process information.

Small Language Model (SLM): a SLM is small relative to a LLM in number of parameters and computational complexity.

Generative AI:? Is a type of AI commonly powered by LLM(s) which can generate content based on art, images, and text.

Hallucination:? Hallucination is possible when AI models are not accurate as a result of flaws in data or training. Avoiding hallucination in models is done through data cleaning and accurate data, which improves results.

Optical Character Recognition (OCR):? OCR refers to extracting text from images for data mining.

Prompt Engineering:? Is the art of writing prompts in such a way that it gives a useful response.

Token:? When you feed a query, it breaks down the text into tokens, which are processed by the AI engine.? Tokens are roughly four characters of text and one to two sentences are around 30 tokens.

Training Data:? Is the data that a program uses to learn and execute its function.

My hope was to give you a quick and short explanation of what I think will assist you in your next conversation about AI.? If there is something which you think could assist others, please comment them below.