Foundations of Artificial Intelligence

 Foundations of Artificial Intelligence

A.     Philosophy:

·       Can formal rules be used to draw valid conclusions?

Yes, formal rules, like in math or logic, can help us make valid conclusions. Think of it like following a recipe: if you follow the steps correctly, you’ll get the right result. In logic, we use rules to make sure our conclusions are correct based on the information we start with.

·       How does the mind come from a physical brain? Where does knowledge come from?

The mind is connected to the brain, but it’s still unclear exactly how the brain creates the mind. Some believe the mind is just the brain working in a certain way, while others think the mind might be something more. Knowledge comes from our experiences, like things we see, hear, or learn, and our ability to think about those experiences.

·       How does knowledge lead to action?

When we know something, it helps us decide what to do. For example, if we know fire is dangerous, we won’t touch it. Knowledge helps us take actions that are smart and safe.

·       Aristotle’s Syllogism:

Aristotle was one of the first to create a clear system of rules for logical thinking. He developed something called syllogisms, which are a way to reason step by step to reach a conclusion.

For example:

·       All dogs are animals (first statement)

·       All animals have four legs (second statement)

·       Therefore, all dogs have four legs (conclusion)

If the two first statements are true, the conclusion must also be true. Aristotle believed this kind of reasoning helped us understand the world.

·       Descartes and Rationalism:

Descartes believed that using our ability to reason (think logically) was very important for understanding the world. This way of thinking is called rationalism. Descartes thought that reason could help us figure out what’s true about the world and ourselves.

B.     Mathematics

·       What are the formal rules to draw valid conclusions? What can be computed?

Formal rules are like steps in a process that we follow to make sure our conclusions are correct. For example, in logic, we follow rules to make sure our reasoning is valid. In terms of computing, there are specific problems and functions that computers can solve, but not everything can be computed.

·       How do we reason with uncertain information?

When we don't have all the facts or we’re not completely sure about something, we use probability or estimates to reason. For example, if you don’t know the weather exactly, you might say, "There’s a 70% chance of rain," based on what you do know.

·       Formal representation and proof algorithms: Propositional logic

In mathematics and logic, propositional logic is a way of using symbols to represent statements that can be true or false. We use rules to combine these statements and prove things logically. Turing was a mathematician who wanted to figure out which problems can be solved by computers, called computable problems.

·       (Un)decidability

In math and logic, decidability means whether a statement can be proven true or false using a system of rules. Undecidable means there are some true things that we can never prove within the system. For example, "A line can be extended infinitely in both directions" is a true statement, but it’s hard to fully prove it using just certain rules.

·       (In)tractability

When we talk about tractability, we mean whether a problem can be solved in a reasonable amount of time. If solving a problem takes a very long time, especially when the size of the problem gets bigger, we call it intractable. For example, if solving a problem takes longer and longer as it gets more complex, it becomes really hard to solve.

·       Probability: Predicting the future

Probability is a way of predicting the chances of something happening in the future. For example, if you roll a fair die, you know there’s a 1 in 6 chance of landing on any number. We use probability to make educated guesses about things we don’t know for sure yet.

C.    Economics:

·       How should we make decisions to maximize payoff?

To make the best decisions, we need to think about what will give us the greatest benefit or reward. This is often called maximizing payoff—choosing the option that will give us the most value, whether it’s money, happiness, or something else important to us.

·       Economics and how we make choices

Economics is all about understanding how people make choices. We make decisions to get the best outcomes for ourselves, called utility (which just means satisfaction or benefit). For example, if you have to choose between two things, you choose the one that will give you the most satisfaction.

·       Decision theory

Decision theory helps us make the best choices, especially when we don’t know everything for sure. It combines two ideas:

·       Probability theory: This is about understanding the chances of different things happening.

·       Utility theory: This helps us measure how much satisfaction or benefit we get from each possible outcome.

By combining these ideas, decision theory gives us a way to think about our choices when the outcome is uncertain, helping us make smart decisions to get the best result.

D.    Neuroscience

How do brains process information?

Our brains process information by receiving signals from our senses, thinking about them, and then making decisions or actions based on that information. It's like the brain is a super-powerful computer that helps us understand and react to the world around us.

What is neuroscience?

·       Neuroscience is the study of the brain and the nervous system (which includes the brain, spinal cord, and nerves). Scientists in this field try to understand how our brain works and how it controls everything we do, feel, and think.

·       What is the brain made of?

The brain is made up of billions of tiny cells called neurons. There are around 100 billion neurons in your brain, and each one is connected to others. These neurons work together to send signals and process information.

·       Neurons as computational units

Neurons are like tiny computers in your brain. They receive information, process it, and send it to other neurons to help make decisions, store memories, and control actions. They work in a very fast and coordinated way to help your brain function.

E.    E.  Psychology

•      Behaviorism

The behaviorism movement, led by John Watson, focused on studying how animals and humans behave based on what they experience (the stimulus) and how they react to it (the response). Behaviorists only looked at what could be measured directly, like how a rat or pigeon reacts to certain things, but they didn’t focus much on things like thoughts or feelings. Behaviorism helped us learn a lot about animal behavior, but it didn’t do as well in explaining human behavior.

·  Cognitive Psychology

Cognitive psychology looks at the brain like a computer. It thinks that our brain processes information, like how a computer runs programs. Psychologists like John Anderson (1980) believed that we can understand how the brain works by figuring out how it processes information—how it stores, remembers, and makes decisions, like a computer running a detailed program. This way of thinking focuses on how mental functions, like memory and problem-solving, are carried out in the brain.

F.      F.Computer engineering:

• ·       What do we need for artificial intelligence to succeed?

For artificial intelligence (AI) to work, we need two things: intelligence (the ability to think or make decisions) and an artifact (something we create, like a computer). The computer is the main artifact we use to build AI.

• ·       The first operational computer

The first operational computer was called the Heath Robinson, and it was built in 1940 by a team led by Alan Turing. It was made for a special purpose: to help decode secret German messages during World War II.

• ·       The first programmable computer

The first operational programmable computer was the Z-3, created in 1941 by Konrad Zuse in Germany. This computer could be programmed to do different tasks, unlike earlier ones that could only do one thing.

• ·       The first electronic computer

The first electronic computer was called the ABC, and it was built by John Atanasoff and his student Clifford Berry between 1940 and 1942 at Iowa State University. This machine used electricity to work, which made it faster than earlier mechanical computers.

• ·       The first programmable machine

Before computers, there was the first programmable machine, which was a loom (a machine used to weave cloth). It was created in 1805 by Joseph Marie Jacquard. The loom used punched cards to store the pattern for weaving, and this idea of using cards to store instructions became important for the development of computers.

G  G. Control theory and cybernetics

• ·       How can artifacts operate under their own control?

A long time ago, Ktesibios of Alexandria (around 250 B.C.) built the first self-controlling machine: a water clock. This clock had a regulator that kept the water flowing at a constant rate. This invention changed what people thought machines could do because it showed that a machine could control itself to keep things steady.

• ·       Modern control theory

In modern times, control theory helps us design systems that can control themselves to achieve certain goals. One part of control theory, called stochastic optimal control, is about creating systems that do the best job they can over time. This idea is similar to what we aim for in artificial intelligence (AI): designing systems that can behave in the best way possible.

• ·       Tools used in control theory

To make these self-controlling systems work, scientists use tools like calculus (the study of change) and matrix algebra (a way to organize numbers to solve problems). These tools help us understand and design systems that can control themselves.

• ·       AI and logical inference

While control theory focuses on making systems that behave optimally (in the best way), AI researchers use other tools like logical inference and computation to solve different types of problems. These problems include things like understanding language, seeing (vision), and planning ahead—things that control theory doesn’t usually focus on.

H    H .Linguistics

 

• ·       How does language relate to thought?

Language and thought are closely connected, and how we learn and use language helps shape how we think. There have been different ideas about how language works and how we learn it.

• ·       B.F. Skinner’s theory on language learning

In 1957, B.F. Skinner published a book called Verbal Behaviour, where he explained the behaviorist approach to learning language. He believed language learning is like any other behavior we learn, through rewards and repetition.

• ·       Noam Chomsky's criticism

Noam Chomsky, another famous thinker, disagreed with Skinner's theory. Chomsky published his own book called Syntactic Structures, where he argued that Skinner’s theory didn’t explain the creativity people show when they use language. For example, we can create and understand sentences we’ve never heard before, which Skinner’s theory didn’t fully explain.

• ·       Linguistics and AI

Modern linguistics (the study of language) and artificial intelligence (AI) started around the same time and grew together. They intersected in a new field called computational linguistics or natural language processing (NLP), which focuses on teaching computers to understand and use language.

• ·       Understanding language is more complex

Over time, it became clear that understanding language is much more complicated than it seemed in 1957. To understand a sentence, we don’t just need to know the rules of grammar. We also need to understand the subject matter (what the sentence is about) and the context (the situation or background in which the sentence is used).

• ·       Knowledge representation

In AI, knowledge representation is about putting information into a form that computers can understand and use to make decisions. This is closely tied to language and informed by linguistic research, because understanding how language works is key to teaching computers to think and reason.