For many people, changes come and go, so we adapt and we keep going. When someone tells us we have to do something new or a new product comes out, we may grumble about it, but we adapt and life goes on. We forget the stress that we felt in the past and continue on until something new comes along.
A similar thing has been happening with machine learning and artificial intelligence. They have made changes in our lives, many of which we don’t even realize. Even if we do know about them, we rarely stop and think about how they are impacting us.
Machine Learning, So What?
For many people and businesses it is good, if not necessary, to adapt. But it is also good to stop and think about the changes. Big data, machine learning and artificial intelligence will change everything about how we live and how we do business.
They will make many jobs easier, but they will also make many jobs obsolete. Machines can now respond to emails with increasing accuracy and human-like qualities. They can analyze data at faster rates and more precisely than humans.
As machine learning becomes more advanced, machines can learn and adapt to new situations based on past situations, without needing a programmer to write code for every specific situation.
As technology advances, the World Economic Forum predicts that the economy will lose 5 million jobs to machines over the next few years.
This means, that no matter what job your job is, we all need to be paying attention to how machine learning is changing the ways in which we work.
Machine Learning and Artificial Intelligence
Machine learning and artificial intelligence live in the same worlds. They are similar, but also quite different.
In general, the term artificial intelligence is older than the term machine learning. Also, most people consider machine learning to be a part of the artificial intelligence world.
Artificial Intelligence (AI)
Artificial Intelligence has had many definitions over the years, but most agree that AI entails machines that think like humans. It is difficult to know when a machine is thinking because there are many things we do not even understand about the human brain. Translating that into machines is difficult. So, mostly AI involves making machines do things humans can do.
This research started in the 1950s and led to the famous Turing test. Researchers use this test to see if a human can figure out if they are interacting with a machine or another human. If the machine is able to mimic human behavior to such a degree that is can trick a human, then it has passed the test.
In 1956, when the term artificial intelligence became well-known in the scientific community, researchers started thinking about how machines could be more like humans. This lead to current advancements in natural language processing, image recognition, and machine learning.
The term machine learning dates back to 1959. At that time, Arthur Samuel defined it as "the ability to learn without being explicitly programmed." He created a computer program that could improve its performance in the game of checkers by learning from its mistakes.
As data mining became more prevalent in the 1990s, machine learning came back from a state of being forgotten to being front and center, as it was useful in statistics and predictive analytics. Machine learning has become so good at those tasks, that some people believe that it should be separate from artificial intelligence. They make the claim that artificial intelligence processes do not have to incorporate machine learning and machine learning does not have to incorporate all aspects of AI.
While the two areas of research have diverged a bit over time, they both have similarities that will continue to keep them intertwined.
Uses of Machine Learning
As machine learning continues to grow, it touches more and more aspects of our lives.
In terms of both data privacy and personal safety, machine learning is making it easier for companies to adapt to changes in malware codes and protect data. In many agencies, machine learning is also making personal security more accurate, as machines track things that humans may have difficulty noticing with great accuracy.
Algorithms are able to use vast amounts of data to better predict how the market will react to various forces. This system of algorithms has made machines much better at investing than people have been.
When diagnosing patients, research shows that machines can diagnose various conditions sometimes years before doctors can. In one case, computer assisted diagnosis was able to diagnose breast cancer from mammography scan up to a year earlier than doctors. In another case, machines have been able to help predict and avoid hospitalizations for diabetes patients.
With algorithms, companies can personalize marketing to make ads specific to individual users.
Financial services are using machine learning to help spot cases of fraud and money laundering.
Search engines can track what you click, how many pages of results you view, how long you stay on pages, and if you choose new keywords to search for. They use that data to offer better results the next time you search for something.
Many companies are using algorithms to help offer more accurate recommendations to customers about things they can buy, videos to watch, music to listen to, and much more.
We have been hearing about smart cars for some time, but have not seen a lot of progress on their use yet. But many top auto executives, 74% in fact, expect smart cars to be in use by 2025.
The Future of Machine Learning
We have barely scratched the surface of how much machine learning can do. Many believe that it will continue to impact our lives in much more significant ways.
Some believe that machines will be able to use our genetic makeup and lifestyle habits to create personalized healthcare plans. Robots will learn through reinforcement and machines will be able to perform sentiment analysis.
As IBM’s Watson becomes more advanced, researchers have stated we have reached a period of cognitive computing. This means that computers are being created with neural networks that mimic how neurological systems in humans work.
These systems allow for deep learning. In this type of machine learning, algorithms run in many layers simultaneously. The next steps in machine learning are related to deep learning: unsupervised learning and simulation-based learning. In time, machine learning will become even more efficient. Machines will learn at the same performance level, in less time and with less data.
The technical aspects of AI and machine learning may seem a bit too abstract to seem relevant, but there is a need for people to understand these terms and their history, as they continue to influence our lives in subtle, yet important ways. The more we understand, the more we can sustain our humanity and our lives, as machines evolve to me more like us.