What's is machines learnings and how does it work
 |
| What is machine learnings |
What is machine learning?
There is no single definition for machine learning yet. But most researchers formulate it something like this:
Machine learning is the science of making AI learn and act like a human, and have it continually improve its learning and abilities based on the real-world data we provide.
Here's how representatives of leading IT companies and research centers define machine learning:
Nvidia It is the practices of using algorithm to analyze data, studying it, and then determined or predicted something's.
Stanford University: "It is the science of making computers work without explicit programming."
MC Kinsey and Co: “Machines learnings is bases on algorithm that can learnings from database without Relying on Rules-based programming.
University's of Washington's: “Machines learnings algorithm can figured out how to performance importantly tasks themselves by Generalizing from the example they're have.
Carnegie's Mellon University's "The fielding of machine learnings is trying to answered the questioning: 'How can we create computer systemic that Automatically improved with experienced, and what are the fundamentals law's that government all learnings processed?
History of Machine Learning
Dmitry Vetrov, research professor, head of the Center for Deep Learning and Bayesian Methods at the Faculty of Computer Science at the Higher School of Economics, notes: initially, computers were used for problems for which the solution algorithm was known to humans.
And only in recent years has it become clear that they can find a way to solve problems for which there is no solution algorithm or one that is not known to humans.
This is how artificial intelligence in a broad sense and machine learning technologies in particular appeared.
network-based computer;
- That same year, his colleague Arthur Samuel invented the first self-learning checkers program. He first coined the term "machine learning", describing it as the process by which a machine exhibits behavior that it was not originally programmed to do;
- In 1967, the first metric algorithm for data classification was created, which allowed AI to use patterns for recognition and learning;
- In 1997, Deep Blue beat world chess champion Garry Kasparov for the first time;
- In 2006, neural network researcher Geoffrey Hinton coined the term “deep learning
- In 2011, Google Brain was founded , a division of Google that deals with projects in the field of AI;
- In 2012, within another division, Google X Lab, they developed a neural network algorithm for recognizing cats in photos and videos. At the same time, Google launched the Google Prediction API cloud service for machine learning, which analyzes unstructured data;
- In 2014, Facebook (since March 21, 2022, the social network has been banned in Russia by a court decision) developed the DeepFace neural network for recognizing faces in photos and videos. Its algorithm works with 97% accuracy;
- In 2015, Amazon launched Amazon Machine Learning, a machine learning platform, and a few months later Microsoft launched a similar one: Distributed Learning Machine Toolkit.
How machine and deep learning, AI and neural networks are related
Machines learnings is a brand of artificially intelligences AI
Neural networks are a type of machine learning.
Deeply learnings is a types of neural network-based Architectures.
Deep learning also involves the research and development of algorithms for machine learning. In particular, learning to correctly represent data at several levels of abstraction. Over the past ten years, deep learning systems have achieved particular success in areas such as object detection and recognition, text-to-speech conversion, and information retrieval.
What problems does machine learning solve?
With the help of machine learning, AI can analyze data, remember information, make predictions, reproduce ready-made models, and select the most suitable option from those proposed.
Such systems are especially useful where it is necessary to perform huge amounts of calculations: for example, bank scoring (calculating a credit rating), analytics in the field of marketing and statistical research, business planning, demographic research, investments, searching for fake news and fraudulent sites.
Leroy Merlin uses Big Data and Machine Learning to find remaining goods in warehouses.
In marketing and e-commerce, machine learning helps configure services and applications to make personalized recommendations.
The Spotify streaming service uses machine learning to create personalized selections of tracks for each user based on the type of music they listen to.
Today, key research is focused on developing data-efficient machine learning—that is, deep learning systems that can learn more efficiently, with the same performance, in less time, and with less data. Such systems are in demand in personalized healthcare, robot learning with reinforcement, and emotion analysis.
Chinese smart vacuum cleaner maker Ecovacs Robotics has trained its vacuum cleaners to recognize socks, wires and other foreign objects on the floor using a variety of photographs and machine learning.
A “smart” camera based on a Raspberry Pi 3B+ microcomputer, using the TensorFlow Light framework , has learned to recognize a smile and take a photo at exactly that moment, as well as perform voice commands.
