artificial intelligence AI is everywhere these days, but the basics of how this impactful new technology works can be confusing. The two most important areas of AI development are “machine learning” and its sub-area “deep learning.” Here we'll provide a brief rundown of what these two important areas are and how they're contributing to the evolution of automation.
Firstly, what is AI?
It’s worth remembering what AI actually is. Proponents of artificial intelligence say that one day Creating machines that can “think” The human brain is an amazing instrument with computational capabilities. Far beyond the capabilities of any currently available machineSoftware engineers working on AI ultimately want to create machines that can not only do everything humans can intelligently do, but also exceed them.
read more: There are a lot of terms in AI. Here is a glossary of the terms you need to know
Currently, the application of AI in business and government is mainly Equivalent to a predictive algorithmits types Suggest the next song They sell on Spotify or try to sell products similar to the ones you bought I bought it on Amazon last weekBut AI evangelists believe that the technology will eventually be capable of much more complex reasoning and decision-making, and this is where ML and DL come in.
Machine learning explained
Machine learning (ML) is a broad category of artificial intelligence that refers to the process of “teaching” a software program how to make predictions or “decisions.” IBM engineer Jeff Croom explains: explain Machine learning is “very sophisticated statistical analysis,” which Crume says allows the machine to make “data-driven predictions and decisions. The more information you feed into the system, the more accurate the predictions you can make,” he says.
Unlike general programming, machines Designed to complete a very specific taskMachine learning revolves around training algorithms to identify patterns in data on their own, and as mentioned earlier, machine learning encompasses a wide variety of activities.
Deep Learning Explained
Deep Learning teeth Machine learning. This is a subcategory of machine learning mentioned above, and like other forms of ML, it focuses on teaching AI to “think”. Unlike other forms of machine learning, DL aims to let algorithms do much of the work. DL is driven by mathematical models called artificial neural networks (ANNs). These networks try to emulate processes that occur naturally in the human brain, such as decision-making and pattern recognition.
Key Differences Between ML and DL
One of the biggest differences between deep learning and other forms of machine learning is the level of “supervision” given to the machine. In less complex machine learning, the computer can probably Supervised learningThis is the process where humans help machines recognize patterns in labeled structured data, thereby improving their ability to perform predictive analytics.
Machine learning relies on vast amounts of “training data” that is often compiled by humans through data labeling (often The salary is not very goodThis process builds a training dataset that can be fed into an AI algorithm to teach it to identify patterns. For example, a company might train an algorithm to: Recognize a specific car brand by photoThe algorithm is then fed a huge set of photos of that car model that have been hand-labeled by humans, and once training is complete, a “test dataset” is also created to measure the accuracy of the machine's predictive capabilities.
On the other hand, DL is They engage in a process known as “unsupervised learning.”In unsupervised learning, machines use neural networks to learn so-called Unstructured or “raw” data—This is data that hasn't yet been labeled or organized into a database. Companies can use automated algorithms to sift through reams of unorganized data, avoiding a lot of human effort.
How Neural Networks Work
An ANN is made up of what are called “nodes”. According to MITAn ANN can contain “thousands, or even millions” of nodes. These nodes are a bit complicated, but in simple terms, they relay and process information, like the nodes in a human brain. In neural networks, the nodes are arranged in organized formats called “layers.” Thus, a “deep” learning network contains multiple layers of nodes. Information travels through the network and interacts with various environments. This contributes to the machine's decision-making process when prompted by a human.
Another important concept in ANN is the “weight”. One critic Weights are given to synapses in the human brain. Weights are just numbers, but they are distributed throughout an AI's neural network and help determine the final output of the AI system. Weights are information inputs that help tune the neural network to make decisions. MIT Deep Dive About Neural Networks As he explains:
A node assigns a number, called a “weight,” to each input connection. When the network is active, the node receives a different item of data (a different number) on each connection and multiplies it by the associated weight. It then sums the resulting products to produce a single number. If that number is below a threshold, the node does not pass the data to the next layer. If the number exceeds a threshold, the node “fires,” which in today's neural nets typically means sending that number (the sum of the weighted inputs) to all output connections.
Simply put, neural networks are configured to help algorithms draw their own conclusions about the data fed into them. Based on their programming, algorithms identify useful connections among large amounts of data and help humans draw their own conclusions based on that analysis.
Why is machine learning important to AI development?
Machines and deep learning help us train machines to perform predictive and interpretive activities that were previously performed only by humans. This has many advantages, but a clear disadvantage is that these machines can (and, to be honest, will) be used for useful as well as malicious purposes, such as government and private surveillance systems, the continued automation of military and defense activities, etc. But they can also be obviously useful for consumer recommendations and coding, and in the best cases, medical and health research. Like any tool, whether artificial intelligence will have a positive or negative impact on the world depends largely on who uses it.
This article originally appeared on Gizmodo.
