This content originally appeared on DEV Community and was authored by Walker-A-1
Intro
So what is machine learning? Machine learning is the process of teaching machines to do a specific task. Or even have a machine teach itself to do a task! Wouldn't that just lead to a robot uprising? Maybe, but probably not.
Wouldn’t it be Easier to Train a Human?
- Machines are better at processing large amounts of information quickly Yes it does take a while to learn coding and how to teach a machine to do an obvious task for a human, machines were never taught right and wrong or how to ride a bike. But once you are able to train a machine to do a simple task it can think much faster than a human. They are easily scalable as well, if you need another one you can make a copy. Being made of code machine brains can be copied perfectly.
- Machines make less errors and don’t get influenced by biases Training a machine to make no mistakes takes a while, but once it's done a machine will never forget. Machines can not define the meaning of life, but assembling the same car one thousand times is simple once broken down. Machines are hyper specialized at their specific tasks, as such they are specifically programmed and built to accomplish that task. They can do the task ten times, check for errors 100 times, every time. They also didn’t inherit the complex emotions that come from having a flesh body and brain. That means they will not carry their own biases to influence their decisions and make mistakes. All in all machines are very good at doing the task they were made for, and very bad at anything else.
- Flexibility Machines are hyper specialists, but that does not mean that they are incapable of change. If the market changes and the specific task is different then your once reliable machine will not be able to keep up. But have no fear, machines are not as afraid of change as people are. If there is a small change in the input and output of your task all you have to do is retrain your machine with slightly different specifications.
- Safety, Machines can do dangerous jobs and take the risk away from human lives So far I have just been talking about data processing and information management. But machines can handle any sort of jobs they have the body for. Take dangerous jobs lifting heavy objects, dealing with dangerous customers, or even handling radioactive materials! Machines are able to take up jobs that pose a risk to humans or are just plain uncomfortable. ATMs and online banking have made it easier for customers to access their finances and gotten rid of the cashier job that is seemingly always at risk of being robbed in the movies. Machines can handle radioactive materials during a spill and machines have not gotten cancer like humans have after dealing with radioactive materials.
- Security bonus
Along with checking for errors many times a second, machines can also check for risks to security or plain old attacks. Cyber attackers are getting more specialized and covering their tracks better every day. But one little thing out of place that might get missed by a human can be checked over and run through a security test every single time by a machine.
Will This Lead to the Terminator Movie? - No, Automation quote from Leto the 2nd No. “Machines just increase the amount of things we can do without thinking” - Leto II from God Emperor of DUNE. Machines and computers are really advanced tools, they make tasks easier for those who do them. Machines and tools have been around forever and they have never turned around to kill us. There are guns ofcourse, but those are just very efficient killing tools, not something that is just inherently lethal to humans. Computers will never evolve and get angry at their creators.
- History of automation and machine automation. We have used machines to automate tasks since 1943, the Colossus computer was made by the British to break Germain encryption in WWII giving them the upper hand to win the war. Even as early as 27,000 BCE they used sinkers to anchor nets in South Korea. Making it so one person did not have to hold a net down the whole time but they could set up multiple fish nets and do their job much more efficiently.
- Automation will not make killer robots, but it will enable rich people to make more money Supporting information All in all machines are not thinking beings and they can not make the decision to murder people. But there is the possibility of making a machine that is more efficient than a gun and of course machines to make rich people richer. How do you Train a Machine? What kind of AI learning algorithms are there? And what are they good for? Supervised learning models take in a lot of labeled data to come to the conclusion that is already known. First it has to make its own guess then evaluate its performance based on the correct answer, every time. This model would be good for object recognition, say if you want a machine to tell you if an image contains a dog or not. At first it has no idea what a dog is, but then seeing many images of dogs it starts to learn to recognize dogs. Does the image have ears? Is it fluffy? Does it have a tail? Pros: this type has very high accuracy and it is easy to evaluate. Cons: This type requires a lot of labeled data, that requires a human to go through a lot of data and label it. Which can be time consuming and expensive. Reinforcement learning models are somewhat what they sound like. You let a machine explore and interact with its environment and give it rewards and penalties based on its actions. This model has been used to teach machines to play video games. Insert your AI into a virtual environment, let it play around, and have very specific parameters for rewards and penalties. This one does take longer to learn its task but once it does, it will be very good at it. Pros: Can handle learning complex behaviors that may even be difficult to teach to a human. Can handle dynamic environments and may be used for decision making. Cons: Needs careful tuning of hyper parameters. Does not generalize task well for new environments.
Conclusion
In conclusion, machine learning can be a great option for automating certain tasks. And no, machines will not make the decision to eliminate humans like in the movie terminator.
Sources:
What jobs are being taken by machines
Five machine learning types
10 most popular AI models
How do machines learn?
Machine training snipette
How did Skynet come to have automated
This content originally appeared on DEV Community and was authored by Walker-A-1
Print
Share
Comment
Cite
Upload
Translate
Updates
There are no updates yet.
Click the Upload button above to add an update.
APA
MLA
Walker-A-1 | Sciencx (2024-06-17T01:44:49+00:00) An intro to Machine Learning. Retrieved from https://www.scien.cx/2024/06/17/an-intro-to-machine-learning/
" » An intro to Machine Learning." Walker-A-1 | Sciencx - Monday June 17, 2024, https://www.scien.cx/2024/06/17/an-intro-to-machine-learning/
HARVARDWalker-A-1 | Sciencx Monday June 17, 2024 » An intro to Machine Learning., viewed ,<https://www.scien.cx/2024/06/17/an-intro-to-machine-learning/>
VANCOUVERWalker-A-1 | Sciencx - » An intro to Machine Learning. [Internet]. [Accessed ]. Available from: https://www.scien.cx/2024/06/17/an-intro-to-machine-learning/
CHICAGO" » An intro to Machine Learning." Walker-A-1 | Sciencx - Accessed . https://www.scien.cx/2024/06/17/an-intro-to-machine-learning/
IEEE" » An intro to Machine Learning." Walker-A-1 | Sciencx [Online]. Available: https://www.scien.cx/2024/06/17/an-intro-to-machine-learning/. [Accessed: ]
rf:citation » An intro to Machine Learning | Walker-A-1 | Sciencx | https://www.scien.cx/2024/06/17/an-intro-to-machine-learning/ |
Please log in to upload a file.
There are no updates yet.
Click the Upload button above to add an update.