What is machine learning? Everything you need to know
What is machine learning? Everything you need to know
Machine learning enables computers to deal with tasks that until now have been performed only by people.
From driving cars to translation speeches, machine learning is driving an explosion in the capabilities of artificial intelligence - messy and unpredictable real-world experiences with the help of software.
But what exactly is machine learning and what is the current boom in machine learning possible?
What is machine Learning ?
At a very high level, machine learning is the process of teaching a computer system how to accurately predict the data fed.
Those predictions could be answered by whether a piece of fruit in a photo is a banana or an apple, with people crossing the road in front of a self-driving car, using the word book in a sentence that relates to a paperback Whether or not hotel reservation, whether an email is spam, or correctly identifies the speech to generate a caption for a YouTube video.
The key difference from traditional computer software is that a human developer has not written code that instructs the system how to tell the difference between a banana and an apple.
WHAT IS THE DIFFERENCE BETWEEN AI AND MACHINE LEARNING?
Machine learning has enjoyed success as of late, but this is just one way to achieve artificial intelligence.
At the birth of the field of AI in the 1950s, AI was defined as a machine capable of performing any task that usually required human intelligence.
AI systems will typically exhibit at least some traits: planning, teaching, logic, problem solving, knowledge representation, perception, motion and manipulation and, to a lesser extent, social intelligence and creativity.
Along with machine learning, there are various other approaches used to build AI systems, including evolutionary computations, where algorithms undergo random solutions and combinations between generations of optimal solutions, and "develop" expert systems. In the attempt, where computers are programmed according to the rules. To mimic the behavior of a human expert in a specific domain, for example the autopilot system flying an airplane.
WHAT ARE THE MAIN TYPES OF MACHINE LEARNING?
Machine learning is generally divided into two main categories: supervised and unhelpful learning.
WHAT IS SUPERVISED LEARNING?
During training for supervised learning, the system is exposed to large amounts of labelled data, for example images of handwritten data to indicate what numbers they correspond to. Given enough examples, a supervised learning system will learn to recognise the groups of pixels and shapes associated with each number and will eventually be able to identify handwritten numbers, which can make a considerable difference between numbers 9 and 4 or 6 and 8 Will be.
While these systems typically require massive amounts of label data to be trained, with some systems millions of instances need to be exposed so that a task can be completed.
WHAT IS UNSUPERVISED LEARNING?
In contrast, the unpublished learning task algorithm, with identifying patterns in the data, is trying to spot similarities that divide those data into categories.
An example might be Air BnB linking together homes available for rent by neighborhood, or Google News creating groups of stories on similar topics each day.
The algorithm is not designed to single out specific types of data, it simply looks for data that can be grouped for its similarity or grouping inconsistencies.
WHAT IS SEMI-SUPERVISED LEARNING?
The importance of vast sets of labelled data for training machine-learning systems may diminish over time, due to the rise of semi-supervised learning.
As the name suggests, the approach combines supervised and unhelpful learning. The technique relies on using a small amount of labelled data and a large amount of unlabelled data for train systems. Labelled data is used partly to train machine-learning models, and then partially trained models are used to label unleaded data, called pseudo-labelling. The model is then trained on the resulting mixture of labelled and pseudo-labelled data.
The feasibility of semi-supervised learning has recently been enhanced by generative advanced network (GAN), machine-learning systems that can use label data to generate entirely new data, e.g. Pokemon from existing images Creating new images of, which may in turn occur. Machine-learning mode used to help train.
WHAT IS REINFORCEMENT LEARNING?
One way to understand reinforcement learning is to think about how someone might learn to play an old school computer game for the first time when they are not familiar with the rules or how to control the game. While they may be a complete novice, ultimately, by looking at the relationship between the buttons they press, what happens on screen and their in-game score, their performance will get better and better.
An example of reinforcement learning is Google Deep Mind's Deep Q-Network, which has defeated humans in a wide range of old video games. The system is fed pixels from each game and determines various information about the state of the game, such as the distance between objects on the screen. It is then assumed that the state of the game and the actions that take place in the game are related to the score it receives.
HOW TO EVALUATE MACHINE-LEARNING MODELS?
Once the training of the model is completed, the model is evaluated using the remaining data that was not used during the training, which helps to measure its real-world performance.
To further improve performance, training parameters can be tuned. An example might be the extent to which the "load" is changed at each step in the training process.
WHY IS MACHINE LEARNING SO SUCCESSFUL?
While machine learning is not a new technology, interest in the field has increased in recent years.
This revival comes on the back of a series of successes, with intensive learning setting new records for accuracy in areas such as speech and language recognition, and computer vision.
To make these successes possible, there are mainly two factors, one is the large amount of images, speech, video, and text, accessible to researchers looking at machine-learning systems.
But even more importantly, courtesy of modern graphics processing units (GPUs) is the availability of large amounts of parallel-processing power, which can be combined together in groups to form machine-learning powerhouses.
Comments
Post a Comment