UNCOMMON MACHINE LEARNING EXAMPLES THAT CHALLENGE WHAT YOU KNOW

摘要： Since machine learning is a relatively new field, the limits of its application are constantly pushed outward. Virtual personal assistants were the stuff of dreams a few years ago, and now they’re seen in every other household. While some examples are conspicuous, here are some ways ML is changing our lives that you may not have thought of.

▲圖片標題(來源：Dataconomy)

Machine learning (ML) is how a system learns and adapts its processes from the patterns found in large amounts of data. When we think of machine learning, some prominent examples come to mind. For instance, the way product recommendations on Amazon are eerily similar to Google searches you’ve done. The scope of machine learning extends far beyond what we know of and see in our daily lives.

EARTHQUAKE DETECTION AND PREDICTION

Machine learning has recently been used to detect seismic waves and analyze the patterns from the data collected and from a million hand-labeled seismograms. The algorithm detects more than twice as many earthquakes as scientists. The idea is that after a sufficient amount of data has been collected and patterns documented, scientists will be able to flag earthquakes in near real-time.

The nature of machine learning implies that it may soon be possible to predict earthquakes before they happen. This has an enormous impact on the planning and preparedness of the disaster management forces, healthcare, firefighters, etc.

BEAUTY INDUSTRY

It is often spoken about how machine learning is used to help consumers test how products would look on their skin from the comfort of their homes. Similarly, producers use this technology to test products while they are still in development. The beauty industry is now using machine learning to create products in a more cost-effective and timely manner.

However, the same technology is now being adapted to ‘objectively’ rank people by beauty. Apps that perform this ranking ask the subject to submit a photo of their faces without makeup and compare the images with those of other people. Some of the factors being considered are symmetry, wrinkles, dark circles, and facial blemishes. This is an uncomfortable notion, not merely because beauty and appearance are sensitive subjects for many people, but also the mental health implications of body dysmorphia and other related issues.

The question arises; why do we need to rank human beings according to external appearance? They say that “beauty is in the eyes of the beholder,” but if we genuinely believe that, then an algorithm to compare and analyze people and their looks is simply uncomfortable. It violates the notion that beauty is subjective and could have terrible effects on peoples’ perceptions of themselves.

SPORTS INJURY PREDICTION

There is a mass of data available in the sports industry used to analyze player performance, success rate, and similar statistics. One of the more exciting uses of machine learning in this field is how sports teams and coaches may use it to predict injuries. The algorithm takes data on muscle movement and observes the regular patterns, which allows teams and coaches to receive an alert when there are anomalies in the default patterns.

This information is beneficial in preventative care for the player. This allows the sports team to save millions of dollars in medical care, lost revenue, and recovery costs. But while every professional athlete is aware of the risk of injury, perhaps having an algorithm (assuming high accuracy) tell you that you will be injured in this period could negatively affect the player’s mental health and performance.

IDENTIFYING WHALES IN THE OCEAN

Underwater noise pollution is a significant threat to marine life such as whales. Recently scientists have learned that machine learning is capable of analyzing acoustic data and accurately detecting whales. Marinexplore and Cornell University had a competition inviting participants to submit their best machine learning algorithm to detect whale call noises from audio recordings. If a whale is identified, cargo carriers, agents, and other interested parties may use the information to plan out shipping routes and prevent collisions. Rerouting ships away from whales would reduce the noise footprints for the whales.

The Cook Inlet beluga whales are endangered, and this machine learning may be what is needed to protect them. Marine life is sensitive to underwater sounds, and the disruptive noises created can cause behavioral changes and physiological harm to the animals. Using algorithms to accurately and efficiently learn the patterns of the whales helps scientists plan out the recovery of this whale population.

DETERMINING MENTAL HEALTH FROM SOCIAL MEDIA

A competition from 9 years ago led to the hypothesis that it is possible to determine psychopathic tendencies in a person from their linguistic patterns and social behavior, both of which are analyzable from Twitter. A similar study was recently conducted on Reddit users. By gathering information from users’ posts on select forums, the algorithm learned to identify people facing difficulties with their mental health.

In the past few years, social media usage has grown immensely since the pandemic rendered in-person interactions impossible. Alongside the increased social media usage, there has been an increasing number of people suffering from poor mental health due to the ongoing pandemic. Developing a means to identify people struggling with their mental health is very useful. Researchers using ML to measure indicators of mental health issues can pass this information on to companies, organizers, and individuals.

There are already apps, like Replika, that chat in a human-like fashion to help users beat loneliness and improve mental health. Imagine a platform that identifies the most at-risk people from their social media messages and then guides them towards apps like Replika to get early help.

THE FUTURE OF MACHINE LEARNING

These examples show us that machine learning is not only dynamic but is also flexible. Its uses extend across industries and even time. Machine learning itself is continuously growing in importance due to the vast amounts of data available and the fact that ML effectively absorbs, analyzes, and makes sense of these large amounts of information. Side by side, the issues that the tech industry may address with ML are of increasing concern. It is crucial that we put our brains to use, discovering new, relevant, and ethical uses for the incredible tool we have, machine learning.

轉貼自： Dataconomy

若喜歡本文，請關注我們的臉書 Please Like our Facebook Page： Big Data In Finance