Machine learning

Machine learning (ML) is a field of study in artificial intelligence. It helps computers learn from data and use that knowledge for new tasks. Instead of being told exactly what to do, these computer programs get better with more information.

A key part of machine learning is deep learning. This uses neural networks — special computer programs — to do even better than older ways. These changes have made many technologies much smarter.

Machine learning is built on statistics and mathematical optimisation. It is also linked to data mining, which looks for patterns in information using methods like unsupervised learning. All of these help computers understand big amounts of data and support many tools we use every day.

History

Relationships to other fields

Artificial intelligence

Machine learning came from the idea of making artificial intelligence (AI). Early AI researchers wanted machines to learn from data, using methods like neural networks and probabilistic reasoning. By the 1980s, AI turned more to expert systems, so machine learning grew on its own. It became its own field in the 1990s, using ideas from statistics and probability theory.

Data compression

Data mining

Machine learning and data mining use similar ways to work but have different aims. Machine learning tries to guess what will happen from known facts. Data mining searches for new patterns in information. Both help each other, with machine learning using data mining skills to get better results.

Generalization

Statistics

Machine learning and statistics are connected but not the same. Statistics draws conclusions from small parts of information. Machine learning finds patterns that can guess new data. Unlike old statistics, machine learning lets the data change the model, using many details to get better.

Statistical physics

Ideas from statistical physics are used to study big machine learning questions, such as learning about deep neural networks. This helps make improvements in areas like medical diagnostics.

Theory

Main articles: Computational learning theory and Statistical learning theory

Machine learning helps computers learn from data. This way, they can make good guesses about new information they haven't seen before. This is called generalisation. Since we can't know everything ahead of time, scientists use probability and math to study how well these learning methods work.

One important idea is to make sure the computer’s model is just right—not too simple and not too complex. If it’s too simple, it might miss important patterns. If it’s too complex, it might only remember the data it was trained on and not work well with new data. Researchers also study how fast these learning methods can work and what kinds of problems they can solve.

Approaches

Machine learning is a way for computers to learn from data. There are three main ways this learning happens:

In supervised learning, the training data is labelled with the expected answers, while in unsupervised learning, the model identifies patterns or structures in unlabelled data.

Supervised learning: The computer gets examples with answers, like learning from a teacher. It tries to find rules that match the answers.
Unsupervised learning: The computer finds patterns in data without any answers given. It looks for groups or structures.
Reinforcement learning: The computer learns by trying things and getting rewards or penalties, like playing a game.

Each method has its own strengths and is used for different kinds of problems. For example, supervised learning is good for predicting things like weather, while reinforcement learning can help a robot learn to move.

Models

A machine learning model is a special kind of mathematical model that can guess what might happen or sort things into groups after learning from information. It learns by looking at examples and getting better over time.

There are many kinds of models used in machine learning. One common kind is called an artificial neural network. This is inspired by how brains work. These networks learn by looking at examples, not by following strict rules. Other models include decision trees, which use branches to help make choices, and support-vector machines, which put data into different groups. Each model is good at different things and is picked depending on what job it needs to do.

Applications

Machine learning has many uses. It helps computers learn from data so they can make decisions or predictions.

Some areas where machine learning is used include agriculture, banking, healthcare, and even space science.

People have used machine learning in fun ways too. For example, a contest was held to improve movie suggestions, and machine learning has helped study art and even fight diseases like COVID-19. It can also help predict weather patterns and make smartphones work better.

Limitations

Machine learning does not always work perfectly. This can happen for many reasons, such as not having enough good information, privacy problems, or using the wrong tools. One challenge is that some machine learning systems can be like a “black box,” meaning even the people who made them can’t always understand how the system makes decisions.

The blue line could be an example of overfitting a linear function due to random noise.

There are also problems like overfitting, where a system learns the information it was given too well and does not work well with new information. Sometimes, machine learning can be tricked into making mistakes by small changes to pictures or data. These challenges show that while machine learning is powerful, it still has limits and needs careful use.

Main article: Explainable artificial intelligence

Main article: Overfitting

Model assessments

To see how well a machine learning model works, we can use different ways to test it.

One way is called the holdout method. We split the data into two parts: most of it helps teach the model, and a smaller part is used to check how well it learned.

Another way is K-fold cross-validation. We divide the data into several small groups. Each group is used, one at a time, to test the model.

We also watch how often the model is right or wrong. Tools like the Receiver Operating Characteristic (ROC) help us know how good a model is, with higher values showing better work.

Ethics

Bias

Main article: Algorithmic bias

Machine learning can sometimes show unfairness because of the data it learns from. If the data has human biases, the machine learning system might copy those biases. For example, a program used for college admissions once treated people with non-European-sounding names unfairly. Some systems have also mislabeled people from certain backgrounds.

Because of these issues, many scientists work to make machine learning fair and helpful for everyone. They think it is important to include people from all backgrounds when making these technologies.

Hardware

Since the 2010s, better computer chips and software have made it easier to teach computers to learn by themselves. Big projects now use special chips called GPUs to do the hard work. These chips help computers learn faster and better.

There are also special chips called Tensor Processing Units (TPUs) made by Google. These TPUs are made just for teaching computers and work very well for big projects. Some computers are built to work more like our brains, which helps them learn in new ways. Machines can also learn on small devices like watches or phones, keeping information safe without needing to send it far away.

Software

Machine learning uses special computer programs to help computers learn from data. There are many tools and programs that people use for this. Some tools are free for anyone to use, while others are made by companies and might cost money.

Here are some of the popular free tools:

Some tools are made by companies but also have free versions:

And here are some tools made by companies that you need to buy:

Journals

Some important journals where researchers share their work on machine learning include the Journal of Machine Learning Research, Machine Learning, Nature Machine Intelligence, Neural Computation, and IEEE Transactions on Pattern Analysis and Machine Intelligence. These journals help scientists learn about new discoveries and ideas in this exciting field.

Conferences

Important conferences where researchers share new ideas about machine learning include the AAAI Conference on Artificial Intelligence, the Association for Computational Linguistics (ACL), and the International Conference on Machine Learning (ICML). Other conferences, like the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD), the Conference on Neural Information Processing Systems (NeurIPS), and the Conference on Knowledge Discovery and Data Mining (KDD), also bring together experts to talk about new discoveries and tools.