5 innovative ways to use machine learning
Hello, Habr! I present to you the translation of the article “5 Innovative Uses for Machine Learning” by Aj Agrawal.
They will come into your life, at least into your business life earlier than you think. Although the time horizon of the coming cannot be accurately predicted, artificial intelligence (AI) promises to fundamentally affect modern society, for better or for worse. The super level (AI) -machine training has received special attention from experts because of the potentially powerful impact on the most important, global industries. Due to the hype, a huge amount of talent and resources flow into this space.
But what exactly is machine learning and why should we care about this in the first place? The answer is that, in the broadest sense, machine learning models of AI applications use independent outcome prediction algorithms. In other words, these models can process gigantic arrays of data, extract conclusions and make accurate predictions without the need for significant human intervention.
Many significant generative consequences result from the accelerated development of this technology, and most of them are ready to significantly simplify the business world.
Here are the five most innovative ways to use machine learning. They will come into your life, at least in your business life, earlier than you expect.
Intensive implementation of autonomous vehicles is a much more efficient form of transportation in the future. Analytical reports suggest that self-driving cars can reduce dangerous traffic (fatal) by as much as 90 percent.
Although we are probably a few years from the start of mass production for the consumer, the adoption of autonomous vehicles is inevitable at this stage in the development of society. However, the time scale for adapting this technology is largely dependent on regulatory actions that often lie outside the control of the technological world.
Software engineers developing these self-propelled “fleets of the future” rely heavily on machine learning technologies to run algorithms that allow vehicles to work autonomously. These models efficiently integrate data from various lidar sensors (a query method using lasers), - radars and cameras - control the vehicle. These well-developed learning algorithms are becoming more intelligent over time, which ensures driving safety.
An important part of the economy, such as the healthcare industry, is still working on inefficient, outdated infrastructure. The main problem points are finding ways to preserve confidential patient information and optimizing the system.
Fortunately, we can use innovative machine learning algorithms (which work without people) to process large amounts of medical data without violating the confidentiality agreement. In addition, we can use these models to better analyze and understand diagnoses, risk factors, and cause-effect ratios.
As Dr. Ed Corbett notes: “It’s clear that machine learning will add another arrow to the quiver of clinical decisions. “Machine learning in medicine is now in the top position,” said Corbett, a health worker at Health Catalyst. Google has developed a machine learning algorithm to help detect cancerous tumors in mammograms. Stanford uses a deep learning algorithm to detect skin cancer. ”
Over the past few years, the international retail sector has consistently generated $ 20 trillion in sales per year. This huge figure covers a gigantic amount of consumer data (demographics, trends, and tastes) made up of an endless stream of trading patterns and trends.
Nevertheless, many retailers are trying to realize the prospects inherent in this valuable information, since information often comes from disparate data warehouses. In the long run, there is a huge opportunity for implementing machine learning models that will allow retailers to better understand their customers and provide a more personalized approach.
Using the previously obtained data, machine learning models can predict everything from which products to recommend to those for which discounts can be launched. Retailers, in particular, can combine digital behaviors to optimize the user's entire journey from the first contact to purchase.
Content moderation is a serious problem for social media platforms such as Facebook and Twitter, in the process of providing reliable information to their audience.
In response to a public outcry against “fake news,” Facebook recently announced that it is hiring 3,000 new employees specifically to keep tabs on the content of the news on the platform. Although this concern extends far beyond social networks, technology conglomerates such as Google are investing heavily in developing their own content monitoring groups to support their fast-growing markets.
Evolving machine learning and AI platforms such as Orions Systems provide proprietary systems for “growing and adapting the interaction between people and artificial intelligence” for tasks such as general moderation of content.
Definitely, these technologies for solving content moderation tasks using innovative tools and resources (for example, analyzing the context and content of each frame from a video) make it possible to increase the productivity of employees. This is an important step forward, which prepares machine algorithms for a very difficult job, the moderation of video materials.
The cost of damage from cybercrime will increase by $ 6 trillion a year by 2021. Experts predict that companies will spend more than $ 1 trillion on cybersecurity services from 2017 to 2021 in order to protect themselves from the growing threat. Most likely, cybersecurity will be a priority, as for startups, and for large enterprises.
Researchers are developing tricky ways to implement machine learning models to detect fraud, prevent phishing, and protect against cyber attacks. Defense systems are trained using the latest data to quickly respond and shield against suspicious activity. Unlike people, these algorithms can work 24 hours a day, seven days a week, without getting tired.
Since these machine learning models have become more accessible to developers, they have steadily begun to gain a large number of endorsements from consumers and enterprises. And when the techno furor happens, it will be interesting to see which techno models have conquered the top.
They will come into your life, at least into your business life earlier than you think. Although the time horizon of the coming cannot be accurately predicted, artificial intelligence (AI) promises to fundamentally affect modern society, for better or for worse. The super level (AI) -machine training has received special attention from experts because of the potentially powerful impact on the most important, global industries. Due to the hype, a huge amount of talent and resources flow into this space.
But what exactly is machine learning and why should we care about this in the first place? The answer is that, in the broadest sense, machine learning models of AI applications use independent outcome prediction algorithms. In other words, these models can process gigantic arrays of data, extract conclusions and make accurate predictions without the need for significant human intervention.
Many significant generative consequences result from the accelerated development of this technology, and most of them are ready to significantly simplify the business world.
Here are the five most innovative ways to use machine learning. They will come into your life, at least in your business life, earlier than you expect.
The widespread use of autonomous vehicles
Intensive implementation of autonomous vehicles is a much more efficient form of transportation in the future. Analytical reports suggest that self-driving cars can reduce dangerous traffic (fatal) by as much as 90 percent.
Although we are probably a few years from the start of mass production for the consumer, the adoption of autonomous vehicles is inevitable at this stage in the development of society. However, the time scale for adapting this technology is largely dependent on regulatory actions that often lie outside the control of the technological world.
Software engineers developing these self-propelled “fleets of the future” rely heavily on machine learning technologies to run algorithms that allow vehicles to work autonomously. These models efficiently integrate data from various lidar sensors (a query method using lasers), - radars and cameras - control the vehicle. These well-developed learning algorithms are becoming more intelligent over time, which ensures driving safety.
More effective healthcare
An important part of the economy, such as the healthcare industry, is still working on inefficient, outdated infrastructure. The main problem points are finding ways to preserve confidential patient information and optimizing the system.
Fortunately, we can use innovative machine learning algorithms (which work without people) to process large amounts of medical data without violating the confidentiality agreement. In addition, we can use these models to better analyze and understand diagnoses, risk factors, and cause-effect ratios.
As Dr. Ed Corbett notes: “It’s clear that machine learning will add another arrow to the quiver of clinical decisions. “Machine learning in medicine is now in the top position,” said Corbett, a health worker at Health Catalyst. Google has developed a machine learning algorithm to help detect cancerous tumors in mammograms. Stanford uses a deep learning algorithm to detect skin cancer. ”
Integrated Retail Management
Over the past few years, the international retail sector has consistently generated $ 20 trillion in sales per year. This huge figure covers a gigantic amount of consumer data (demographics, trends, and tastes) made up of an endless stream of trading patterns and trends.
Nevertheless, many retailers are trying to realize the prospects inherent in this valuable information, since information often comes from disparate data warehouses. In the long run, there is a huge opportunity for implementing machine learning models that will allow retailers to better understand their customers and provide a more personalized approach.
Using the previously obtained data, machine learning models can predict everything from which products to recommend to those for which discounts can be launched. Retailers, in particular, can combine digital behaviors to optimize the user's entire journey from the first contact to purchase.
Content Moderation Optimization
Content moderation is a serious problem for social media platforms such as Facebook and Twitter, in the process of providing reliable information to their audience.
In response to a public outcry against “fake news,” Facebook recently announced that it is hiring 3,000 new employees specifically to keep tabs on the content of the news on the platform. Although this concern extends far beyond social networks, technology conglomerates such as Google are investing heavily in developing their own content monitoring groups to support their fast-growing markets.
Evolving machine learning and AI platforms such as Orions Systems provide proprietary systems for “growing and adapting the interaction between people and artificial intelligence” for tasks such as general moderation of content.
Definitely, these technologies for solving content moderation tasks using innovative tools and resources (for example, analyzing the context and content of each frame from a video) make it possible to increase the productivity of employees. This is an important step forward, which prepares machine algorithms for a very difficult job, the moderation of video materials.
Enhanced Cybersecurity
The cost of damage from cybercrime will increase by $ 6 trillion a year by 2021. Experts predict that companies will spend more than $ 1 trillion on cybersecurity services from 2017 to 2021 in order to protect themselves from the growing threat. Most likely, cybersecurity will be a priority, as for startups, and for large enterprises.
Researchers are developing tricky ways to implement machine learning models to detect fraud, prevent phishing, and protect against cyber attacks. Defense systems are trained using the latest data to quickly respond and shield against suspicious activity. Unlike people, these algorithms can work 24 hours a day, seven days a week, without getting tired.
Since these machine learning models have become more accessible to developers, they have steadily begun to gain a large number of endorsements from consumers and enterprises. And when the techno furor happens, it will be interesting to see which techno models have conquered the top.