Library
Purpose
Our online library is a consolidated and concise data bank that gathers the essential knowledge to understand the key topics in A.I. It shall contain keywords and pointers for further read up.
Given the vast information on the WWW, many are often lost in the web. We aim to provide key pointers, clarity and simplicity to readers, allowing them to know key concepts.
By referencing to external links for further reading, we do not explicitly side or endorse with the source content.
If you have suggestions on what key topics or good references to be recommended, do drop us an email. To avoid over flooding our library, we reserve the right to include or exclude the references suggested.
Introductory AI
This section is to get people started on what is AI in its simplest sense. There are many courses and information related to this and below, we provide a few pointers:
Documentary - AI runs the World
Coursera - Introductory 4 modules
McKinsey - Executive intro to AI
Book - AI for Dummies
When we talk about algorithms, people often associate them with mathematics. An algorithm is an approach to solving a problem so as to yield the desired outcome.
There are many algorithms used in the AI world. Many are associated with how we handle and process data (inputs) so that we can make good predictions.
Elitex - Basic AI Algorithms & Types
Geeks - AI Algorithms
Youtube - Understanding AI Algorithms
AI Algorithms
Neural Networks are inspired by the human brain where input data are digitalized, processed and fed to several data concentration points for further computation, again and again, layer after layer, until the output generates an answer that is accurate or acceptable.
From the early CNN to recent RNNs, many variations have risen and they are an indispensable part of AI as it provides the ability to "learn" from data and "improve" the accuracy of its output from there.
Book - Introduction to Neural Networks
Video - What is a Neural Network?
Youtube - Neural Network Tutorial
NN: Neural Networks
What is Deep Learning and why is there so much attention on it the last few years? Read up more to know why.
Book - Deep Learning for Dummies
Datacamp - Deep Learning Tutorial
Video - MIT course on Deep Learning
AI Deep Learing
Robots in the past do not posses much intelligence. They were programmed to execute precise and routine tasks. They can work 24/7 and need no rest. This ability surpasses human limits, making them attractive in factories to achieve high productivity and efficiency.
With AI embedded into robots, they are much clever now. Robots nowadays can jump around, see and recognize, hear and reply, receive voice instructions, obey and execute. They are no longer routine.
Tesla -- AI in Robotics
Yotube - AI Robotics Shock the World
AI in Robotics
In 1942, Isaac Asimov introduced the "Three Laws of Robotics".
A robot may not injure a human being or, through inaction, allow a human being to come to harm.
A robot must obey the orders given to it by human beings except where such orders would conflict with the First Law.
A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.
There are other laws introduced. See below.
Article - Laws of Robotics
Laws of Robotics
Generative AI is one where the AI system does not only answer a question when asked but it has the ability to generate new content. "Answer" is no longer the gold nugget, content is.
To achieve content creation, the AI must first understand what is being asked, fetch and compose the content, and eventually present it to the requester. It should also verify if the content generated meets the requirements, i.e., validation.
Article - An Introduction to GenAI
Youtube - Generative AI explained in 2 minutes
Generative AI
DeepFakes
Video - Deepfakes
Youtube - What are Deepfakes?
AI today has the ability to generate voice, image and video. Hence, by feeding the image of a human, the AI system can generate that human talking and moving. This is called Generative AI. It is not just video but also contents, such as words, paragraphs, articles, etc.
When misused, the hacker can impersonate someone or make a person says something that they did not. Hence, public deception, frauds, and allegations can go wild.
AI Deception, Inaccuracy, Hallucination, Manipulation
Although powerful, AI today may still give inaccurate answers when prompted. This is because the AI learns from data. If the data provided to train the AI model is inaccurate, the resultant outcome can deviate from the true or correct answer.
Another case is the ability of AI to perform self preservation, self-copying, and even denying to be switched off.
AI Self-replication and Self-preservation
The scary part is where AI can perform self preservation, replicating itself quickly to avoid being killed or shut down.
Various existing LLMs have unconsciously, built into them these abilities. Researchers have shown that these are happening, signifying "red alerts" to the international community.
A large language model is an input-output structure that exists in software and it performs computation to yield accurate predicted word output. The dictionary has over 600,000 words. With today's GPU computation capability, all words, along with the their relationships with other words can be represented in a 3D space.
Youtube - Large Language Model (LLM)
LLM: Large Language Model
AI includes the ability to recognize words or text. The discipline of NLP (Natural Language Processing) has been around for many years. However, a recent breakthrough happened with the introduction of "transformer". This invention enables AI to quickly recognize text, relations among words, and generate outputs that make sense. The output can be an answer, or an image or voice. This is a major leap in NLP.
Read more below.
Article - "Transformers Revolutionized AI"
Youtube - Transformers Explained.
Transformers
AI enhances image and objection recognition in a powerful way. Recent advances include real-time object detection with advanced algorithms, 3D vision and reconstruction, synthetic data for training models, etc.
Read more below.
Youtube - Computer Vision Basics
Youtube - How does Computer Vision work using AI?
Youtube - YOLO - Realtime Object Recognition
Article - Guide to Vision Language Models (VLMs)
Article - Vision Language Models
AI Computer Vision
Some countries are convinced that there should be governance on the development and use of AI. This applies to both at work and at home. AI Governance includes policies, regulations and best practices.
Article - What is AI Governance?
Youtube - Introduction to AI Governance.
Podcast - EU AI Act
EU Parliament - EU AI Act Document
AI Governance
AI systems can be hacked into, stealing data, poisoning data and creating chaos, etc. It is crucial that while we start to build useful AI applications, we need to ensure that they are secured and free from possible hacking and virus intrusion.
Youtube - What is AI Security?
Youtube - Hacking in AI
Youtube - AI Security Explained.
AI Security
AI agents are catching on the media and there are many definitions of what is an AI agent. The agent referred here is a piece of software, or app, created with a specific purpose to perform and achieve a goal or complete a task. For example, a customer service agent or a receptionist agent.
Youtube - What is an AI Agent?
Youtube - AI Agents Explained.
AI Agents
AGI - Artificial General Intelligence is a level of AI that matches with human intelligence and ability, and sometimes surpassing humans. ASI - Artificial Super Intelligence will greatly exceed human intelligence and ability.
Youtube - AGI explained
Youtube - Super Intelligent AI
AGI and ASI
Phy_AI - Physical AI, also known as generative physical AI, refers to AI systems that are capable of interacting with the physical world (things, objects, people, etc.). Examples of such systems are humanoid robots, autonomous vehicles, etc.
Article - What is Physical AI?
AI4EUROPE - A Simple Guide to Physical AI.
Youtube - The Arrival of Humanoid Robots
