What are the machine learning courses

Python course

Difference between "machine learning" and "artificial intelligence"

Andrew Moore, former dean of the School of Computer Science at Carnegie Mellon University: "Artificial intelligence is the science and engineering of making computers behave in ways that, until recently, we thought required human intelligence."

Even the question "What is intelligence?" difficult to answer. "What is Artificial Intelligence?" depends on the answer to the previous question.

The subdivision into

weak AI and strong AI

Strong and weak AI

weak AI:

  • deals with specific application problems
  • Support of human thinking in some areas
  • capable of learning in the sub-area
  • no awareness

strong AI:

  • "general intelligence"
  • Comparable to human intelligence, but does not have to be the same, could be different
  • logical thinking
  • Communication skills, natural language
  • generally capable of learning
  • Awareness?
  • Sentience, emotions?
  • Self-awareness?

Strong AI would be computer systems that work at eye level with people and can support them with difficult tasks. In contrast, weak AI is about mastering specific application problems. Human thinking and technical applications are to be supported here in individual areas. [1] The ability to learn is a key requirement of AI systems and must be an integral part that cannot be added afterwards. A second main criterion is the ability of an AI system to deal with uncertainty and probabilistic information. [2] In particular, those applications are of interest, for the solution of which, according to general understanding, a form of "intelligence" appears to be necessary. Ultimately, the weak AI is about the simulation of intelligent behavior using the tools of mathematics and computer science, it is not about creating consciousness or about a deeper understanding of intelligence. While the creation of strong AI has failed to this day because of its philosophical question, significant progress has been made on the side of weak AI in recent years.

A strong AI system doesn't have to have a lot in common with humans. It will likely have a different cognitive architecture and its developmental stages will also not be comparable to the evolutionary cognitive stages of human thought (evolution of thought). Above all, it cannot be assumed that an artificial intelligence possesses feelings such as love, hate, fear or joy. [3] However, it can simulate behavior corresponding to such feelings.

What is machine learning?

Let's start with a very "old" attempt at a definition by Arthur Samuek, an IBM pioneer:

"Machine Learning: Field of study that gives computers the ability to learn without being explicitly programmed." ("Machine Learning: An area of ​​study in which computers can learn without being explicitly programmed.")

A good attempt, but many questions remain unanswered. Almost 40 years later, in 1998, Tom Mitchell coined a "well-placed learning problem" as follows:

"Well posed Learning Problem: A computer program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T, as measured by P, improves with experience E." A computer program should learn from experience E in relation to a task T and a performance measure P if its performance improves on T, measured by P, with experience E.)

(Note: A mathematical problem is called correctly posed (also well posed, well posed or appropriately posed) if the following conditions are met:

  • The problem has a solution (existence).
  • This solution is uniquely determined (uniqueness).
  • This solution always depends on the input data (stability). )

So what is machine learning?

Machine learning is the process of automatically extracting knowledge from data, usually with the aim of making predictions about new, invisible data. As already mentioned, a spam filter could be implemented using a classifier based on machine learning.

At the heart of machine learning is the concept of automating decision-making from data without the user specifying explicit rules as to how this decision should be made. In the case of e-mail, the user does not provide a list of words or characteristics that make an e-mail spam. Instead, the user provides examples of spam and non-spam emails that are marked as such. This is then the so-called learning set.

The goal of a machine learning model is to predict new, previously invisible data. In a real application, we are not interested in marking an email that has already been marked as spam or not. Instead, we want to make life easier for the user by automatically classifying new incoming email.

These examples are then learned or trained by the algorithm.