Is the essence of machine learning reduced to the classification of the analyzed data?

Question

Machine learning

I watch and read various materials on machine learning, everyone speaks complicated, in mathematical terms and talks a lot, but they always come to the same point - the machine must recognize. Did I understand correctly that today the task is to teach machines to accurately identify (better than a person) various forms of something (the subject of pictures, people and animals, predict the behavior of something or someone) based on the collected data?

Answer 1

Typology of machine learning tasks - I think this article best reveals the essence of machine learning in terms of the tasks to be solved.

Answer 2

It seems to me that the Russian Wikipedia has a basic set of tasks.
So, the examples are not from the classification.
1. We write a search engine. Found 100500 pages, you need to highlight the top 10 most relevant. Ranking problem.
2. We are writing a service for comparing prices for apartments. We have a market database. Another apartment arrives with its own parameters: house, area, distance to the metro, local infrastructure. We need to restore the price. Regression.
3. We maintain a popular Instagram account. We have an audience, we want to understand its structure, think about what can be sold there. It is required to select the communities into which subscribers are divided (usually they are subscribed to each other). clustering problem.
4. We are film search. We have a huge sparse matrix (users x movies). Who liked what and how much. We want to learn how to advise whom to recommend which film. recommender systems.

Answer 3

www.machinelearning.ru/wiki/index.php?title=%D0%9C...

Answer 4

The essence of the machine is not to develop or code algorithms for specific tasks.

Answer 5

Machine learning - selecting a function and its parameters based on data to best transform new data into a response. Machine learning is a machine implementation of decision making. I know the story - what will happen next? I see a picture - what's on it? I do not know the characteristics of the film - to whom should I recommend it? But it all started with a search: if the user entered a set of words - what is the ranked list of sites for him? Basically, this is an optimization problem. Finding a minimum for the function of discrepancy between prediction and reality. It seems that this is how our brain works.

Answer 6

The main task of machine learning is to develop an artificial intelligence that has a perception at least the same as that of a person in order to automatically analyze the facts obtained using this perception and predict the future based on this.
Ideally, AI should replace humans in all areas - because of this, many professions may soon become completely unnecessary. The last ones left are the AI ​​programmers. Already, people involved in machine learning have the highest annual income in the IT world. The competition is not very big yet, so if you start learning ML, you can be in time.
There are a lot of machine learning courses on Coursera, but maybe. still need books on linear algebra, probability theory, statistics and mathematical analysis.

Answer 7

Well, almost.
Example:
The tic-tac-toe bot, if he managed to win, will write down the combination to himself. If not, then he will throw it to the player, and see how he solves it