{"id":17734,"date":"2024-09-11T15:08:23","date_gmt":"2024-09-11T15:08:23","guid":{"rendered":"https:\/\/surreyborder.org.uk\/?p=17734"},"modified":"2024-10-26T21:54:36","modified_gmt":"2024-10-26T21:54:36","slug":"a-simple-guide-to-ai-terminology","status":"publish","type":"post","link":"https:\/\/surreyborder.org.uk\/index.php\/2024\/09\/11\/a-simple-guide-to-ai-terminology\/","title":{"rendered":"A simple Guide to Ai Terminology"},"content":{"rendered":"\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t

As Ai develops and changes,\u00a0 it’s very easy to become lost or left behind.
Certainly, it’s reasonable to believe that the world is largely totally unprepared for what is about to occur.
With that in mind, it’s probably useful to get a basic understanding of the Ai terminology that will soon become common.
This is my plain English translation of the jargon.<\/p>

Welcome to the future.
Jim Reed<\/h5>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t

Core Ai Concepts<\/span><\/h3>

AI (Artificial Intelligence):<\/strong><\/span>
Computer systems that can perform tasks that typically require human intelligence, like understanding language, recognizing images, or solving problems. They’re “narrow” AI systems designed for specific tasks.<\/p>

AGI (Artificial General Intelligence):<\/span><\/strong>
A hypothetical future AI system that could match human-level intelligence across all tasks and could learn new skills just like humans do. Like having a computer that’s as adaptable and capable as a human mind.<\/p>

ASI (Artificial Superintelligence):<\/span><\/strong>
A theoretical future AI system that would be far smarter than humans in virtually every field. Like having a mind that’s to human intelligence what human intelligence is to ant intelligence.<\/p>

Safe ASI:<\/span><\/strong>
The concept of developing superintelligent AI systems with reliable safeguards to ensure they remain beneficial and aligned with human values and wellbeing. Like creating an incredibly powerful tool that’s guaranteed to help rather than harm humanity.<\/p>

Agents:<\/span><\/strong>
AI systems that can independently perform tasks, make decisions, and interact with their environment to achieve specific goals. Like having a virtual assistant that can actually complete tasks on its own rather than just answering questions.<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t

Learning and Training Terms<\/span><\/h3>

Machine Learning:<\/span><\/strong>
A way for computers to learn from examples rather than following strict rules. It’s like how you learn to recognize cats after seeing many cat pictures, rather than memorizing a list of cat features.<\/p>

Deep Learning:<\/span><\/strong>
A more complex type of machine learning using many layers of neural networks. It’s particularly good at handling complicated tasks like understanding speech or images.<\/p>

Training Data:<\/span><\/strong>
The examples we feed to AI to help it learn – like showing a child many pictures while teaching them words.<\/p>

Supervised Learning:<\/span><\/strong>
Teaching AI by showing it examples along with the correct answers, like a teacher grading homework.<\/p>

Unsupervised Learning:<\/strong><\/span>
Letting AI find patterns in data on its own, without telling it what to look for. Like letting someone group similar objects without telling them how to categorize them.<\/p>

Reinforcement Learning:<\/strong><\/span>
Teaching AI through trial and error with rewards for good outcomes. Like training a dog with treats when it does something right.<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t

Ai Behavior and Performance<\/span><\/h3>

Bias:<\/span><\/strong>
When an AI makes unfair or incorrect decisions because of problems in its training data. Like having a teacher who only shows you examples from one perspective.<\/p>

Overfitting:<\/span><\/strong>
When an AI learns its training data too perfectly and can’t handle new situations well. Like memorizing test answers without understanding the subject.<\/p>

Hallucination:<\/span><\/strong>
When AI makes up information that sounds plausible but isn’t true. Like a student confidently giving a wrong answer while sounding convincing.<\/p>

Dataset:<\/span><\/strong>
The collection of information used to train an AI. Like a library of examples the AI learns from.<\/p>

Embedding:<\/span><\/strong>
Converting words, images, or other data into numbers that AI can understand and work with. Like translating everything into a universal language for computers.<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t

Training Concepts<\/span><\/h3>

Batch Size:<\/span><\/strong>
The number of examples an AI looks at before updating what it’s learned. Like giving a student several practice problems before checking their work.<\/p>

Epoch:<\/span><\/strong>
One complete pass through all the training data when teaching an AI. Like going through an entire textbook once.<\/p>

Foundation Model:<\/span><\/strong>
A large AI system trained on massive amounts of general data that can be adapted for specific tasks. Think of it as a basic education that can be built upon for different careers.<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t

Additional Technical Concepts<\/span><\/h3>

Attention Mechanism:<\/span><\/strong>
The way AI focuses on different parts of information to understand context. Like knowing which parts of a sentence are most important for answering a question.<\/p>

Dropout:<\/span><\/strong>
A technique that randomly turns off parts of an AI during training to make it more reliable. Like practicing solving problems with some hints hidden.<\/p>

Vector<\/span><\/strong>:
A way of representing words, images, or concepts as a list of numbers that AI can understand. Like giving everything a unique digital fingerprint.<\/p>

Sequence-to-sequence:<\/span><\/strong>
AI models that convert one sequence into another, like translation. Similar to reading something in one language and writing it in another.<\/p>

Edge AI:<\/span><\/strong>
AI systems that run on local devices rather than in the cloud. Like having a smart assistant that works without internet connection.<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t

\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t

Common Ai Terms<\/span><\/h3>

Large Language Models (LLMs):<\/span><\/strong>
Computer programs that understand and generate human language after being trained on huge amounts of text. They’re like very advanced autocomplete systems that can write, answer questions, and hold conversations.<\/p>

Neural Networks:<\/span><\/strong>
Computer systems inspired by how human brains work, with interconnected “nodes” that learn patterns from data. They’re the basic building blocks of modern AI.<\/p>

Synthetic Data:<\/strong><\/span>
Artificially created information that mimics real data, used to train AI when real data is scarce or sensitive. It’s like creating realistic practice scenarios.<\/p>

Inference:<\/span><\/strong>
The process when an AI actually makes decisions or generates outputs using what it learned during training. It’s similar to how you use knowledge you’ve learned to answer questions on a test.<\/p>

Distillation:<\/span><\/strong>
Making a smaller, faster AI model by having it learn from a larger model, like creating a compact “cliff notes” version that maintains most of the important capabilities.<\/p>

Simulations:<\/span><\/strong>
Virtual environments where AI can practice and learn tasks safely before working in the real world. Think of it as a simulator game, but for training AI.<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t

Technical Terms<\/span><\/h3>

Weights:<\/span><\/strong>
The strength of connections between different parts of a neural network, which the AI adjusts as it learns. Like how your brain strengthens connections between neurons when learning something new.<\/p>

SLM (Small Language Model):<\/span><\/strong>
A more compact version of language AI that’s designed to be faster and use less computing power than larger models. Like having a pocket dictionary versus a full encyclopedia.<\/p>

Back Propagation:<\/span><\/strong>
The way neural networks learn from their mistakes by adjusting their weights backwards through the system. Like tracking back through your work to find where you made a mistake and fixing it.<\/p>

Parameters:<\/span><\/strong>
The adjustable “knobs” inside an AI model that it uses to make decisions. The more parameters, generally the more complex things the AI can understand.<\/p>

Tokenization:<\/span><\/strong>
Breaking text into smaller pieces (tokens) that the AI can process. Similar to breaking a sentence into words and punctuation marks.<\/p>

Fine-tuning:<\/span><\/strong>
Adapting an AI model for a specific task by giving it additional training on specialized data. Like taking a general education and then specializing in a particular field.<\/p>

Prompt:<\/strong><\/span>
The input or instruction given to an AI to get it to do something. Think of it as the question or task you’re giving the AI to work on.<\/p>

Transformer:<\/span><\/strong>
A specific design for AI models that’s particularly good at understanding context in language. It’s like having a super-powered reading comprehension system.<\/p>

Latency:<\/span><\/strong>
How long it takes for an AI to respond after you give it a task. Like the time between asking a question and getting an answer.<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t

Advanced Concepts<\/span><\/h3>

API (Application Programming Interface):<\/span><\/strong>
A way for people and programs to interact with AI systems. Like having a universal remote control that works with different devices.<\/p>

Confidence Score:<\/strong><\/span>
How sure an AI is about its answer. Similar to how confident you feel about knowing the correct answer to a question.<\/p>

Transfer Learning:<\/span><\/strong>
When an AI uses knowledge learned from one task to help with a different task. Like using your knowledge of Spanish to help learn Italian.<\/p>

Zero-shot Learning:<\/span><\/strong>
When AI can handle tasks it wasn’t specifically trained for. Like being able to answer a test question using general knowledge rather than studying that exact topic.<\/p>

Few-shot Learning:<\/span><\/strong>
When AI learns from just a few examples. Like quickly learning a new game after seeing it played only once or twice.<\/p>

Multimodal AI:<\/span><\/strong>
AI systems that can work with different types of input like text, images, and sound together. Like having someone who can understand both spoken words and sign language.<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t

Advanced Applications<\/span><\/h3>

Model Compression:<\/span><\/strong>
Techniques to make AI models smaller and faster while keeping most of their abilities. Like making a shorter version of a book that keeps the main ideas.<\/p>

Ensemble Learning:<\/span><\/strong>
Using multiple AI models together to get better results. Like getting opinions from several experts before making a decision.<\/p>

Semantic Search:<\/span><\/strong>
Finding information based on meaning rather than just matching exact words. Like understanding that a query about “cars” should also find information about “automobiles.”<\/p>

Data Augmentation:<\/span><\/strong>
Creating variations of training data to help AI learn better. Like showing the same picture from different angles to help someone recognize an object.<\/p>

Prompt Engineering:<\/span><\/strong>
The skill of writing effective instructions for AI to get better results. Like knowing exactly how to phrase a question to get the most helpful answer.<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t<\/i>\t\t\t<\/span>\n\t\t\t\t\t\t\t\t\tAi Home Page<\/span>\n\t\t\t\t\t<\/span>\n\t\t\t\t\t<\/a>\n\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"","protected":false},"author":12,"featured_media":17004,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"iawp_total_views":20,"footnotes":""},"categories":[42],"tags":[],"class_list":["post-17734","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai"],"_links":{"self":[{"href":"https:\/\/surreyborder.org.uk\/index.php\/wp-json\/wp\/v2\/posts\/17734","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/surreyborder.org.uk\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/surreyborder.org.uk\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/surreyborder.org.uk\/index.php\/wp-json\/wp\/v2\/users\/12"}],"replies":[{"embeddable":true,"href":"https:\/\/surreyborder.org.uk\/index.php\/wp-json\/wp\/v2\/comments?post=17734"}],"version-history":[{"count":34,"href":"https:\/\/surreyborder.org.uk\/index.php\/wp-json\/wp\/v2\/posts\/17734\/revisions"}],"predecessor-version":[{"id":17857,"href":"https:\/\/surreyborder.org.uk\/index.php\/wp-json\/wp\/v2\/posts\/17734\/revisions\/17857"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/surreyborder.org.uk\/index.php\/wp-json\/wp\/v2\/media\/17004"}],"wp:attachment":[{"href":"https:\/\/surreyborder.org.uk\/index.php\/wp-json\/wp\/v2\/media?parent=17734"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/surreyborder.org.uk\/index.php\/wp-json\/wp\/v2\/categories?post=17734"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/surreyborder.org.uk\/index.php\/wp-json\/wp\/v2\/tags?post=17734"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}