Algorithms tell the world what we think, discern our moods

Moods hedonometer algorithms
Happiness, elusive or not, can now be measured by computer algorithms that analyze our moods, called hedonometer. Credit: Wikimedia Commons / CC BY-SA 2.0

A machine called Hedonometer – yes, you read that right – now discerns how happy we are with the words and phrases we use online, via algorithms that target our moods.

Distilling monumental amounts of data, the system can tell how world events impact society in multiple ways. But it also shows us how much our privacy has deteriorated over the years, as our words have passed through some sort of public mill to be spat out on the other end.

Grecian Delight supports Greece

Chris Danforth, more than anyone else, is responsible for the perfection of this mega machine that analyzes our thoughts and reflections online and publishes reports on how we feel every day. A graduate of the University of Vermont, the professor of applied mathematics devotes much of his time to interpreting the results of the Hedonometer, which itself is the product of decades of research at the University.

Operating 24 hours a day, 7 days a week, the Hedonometer analyzes some 50 million tweets on Twitter every day and then draws up an overall picture of the mood of the public.

Using computers to assess the emotional tone of words has been the goal of researchers for as many decades as there have been computers. But in his construction of the Hedonometer, Danforth had to teach a machine how to understand the emotions behind Tweets.

Sentiment analysis was perfect for March 2020

The process by which the machine understands emotions is called “sentiment analysis”.

Unsurprisingly, 2020 has been the saddest year for all of us, if the hedonometer is any measure of it, and it is – and it wasn’t even close.

The machine, which has been in use since 2008, got its biggest drive in March of last year as the pandemic hit the world head-on, causing fear, unease and even panic.

As reported in Knowable magazine, Wednesday March 11 was the culmination of this arc of bad news, and it was reflected in the public discourse. It was then that the NBA suspended its season; President Trump has issued an order suspending certain travel from Europe to the United States; and popular actor Tom Hanks announced that he and his wife, Rita Wilson, had contracted the virus.

The very next day, Danforth and his team noticed that sentiment analysis showed that there had been a sharp drop in national mood. It also lasted for weeks, as Americans and others began to realize that the pandemic would turn their social lives upside down as well as the way they worked, studied and traveled.

“In the history of our instrument, over a decade, we have never seen an event that actually lingers in our collective mood for more than a day or two,” Danforth said in an interview. “Since March 12, the mood (was) dramatically depressed on Twitter.”

The use of machines to assess mood is a tool that naturally lends itself to many fields of endeavor, including marketing, medical research and other types of research, as well as the news media.

In addition to these more mundane areas where assessing audience sentiment is very beneficial, the hedonometer can also be used for more ethereal purposes, including showing how sadder a minor chord in music is than a major chord. .

Naturally, marketing information is sought after by businesses who can use things like reviews on Yelp to determine exactly who to target in their campaigns.

Employers Use Algorithms to Judge Internal Employee Writings Using Algorithms

A little more worryingly, employers can use it to gauge their employees’ moods on their own internal social networks and correspondence. More usefully, perhaps, the technology can be used to identify depressed people who may need immediate medical intervention.

Before, assessing public sentiment was a much trickier business. As Danforth puts it: “In the social sciences we tend to measure easy things, like gross domestic product. Happiness is an important thing that is difficult to measure.

Computers are notoriously bad at deconstructing human language to exploit it for emotional clues. However, there are many clues to the emotions behind any written text that computers can take note of, even when they don’t understand the meaning of the words.

One obvious way for the system to calculate our moods is to count certain words and see how they stack up against each other, with positive words weighed against negative terms. It’s even more effective to weight each word according to the amount of feelings behind it.

Of course, the word “Excellent” is more powerful than “good”. This is where humans come in – the Danforth team assigns these “weights” to words, which are part of emotional word dictionaries, called lexicons.

These lexicons have now become the basis of almost all sentiment analysis.

Of course, human language is infinitely complex, and things like humor – especially sarcasm – just ignore it.

Other issues, such as word order, are also important in assessing our emotions. Just counting the number of positive words versus negative words doesn’t work, as can be seen in the following sentence: “I’m so glad the weather wasn’t terrible and stuffy like it was last week, this which was unbearable.

Obviously, the person is happy right now – but the number of negative words, such as terrible, suffocating, and unbearable, exceeds the number of positive words. Context matters.

To avoid this problem, there are now machine learning algorithms that train computers to recognize patterns, including essential relationships between words. For example, you can tell the computer that pairs of words, such as “shore” and “river” are often seen together. However, if “bank” and “money” appear in the same sentence, the author is probably referring to a different type of bank.

Neural networks are advanced types of algorithms

Word integrations, developed by Tomas Mikolov of Google Brain in 2013, were another major advancement in the tools used to discern moods and meanings..

In this method, each word is converted into a list of 50 to 300 numbers, called a vector. These numbers are like indelible fingerprints that describe not only each term, but also the other words it is commonly associated with.

This tool is part of what researchers call a neural network method. The more layers are added to these networks, the more information can be gleaned from our publications and writings of all kinds.

The first sensitive scan achieved an accuracy of about 74 percent. Today’s ultra-sophisticated neural networks operate with greater than 94% accuracy, which comes close to that of a human being in their ability to judge emotions, according to To know.

The University of Vermont hedonometer still uses a lexicon, and Danforth has no plans to change this method, which could be more related to the time it takes to “train” a computer to use networks. neuronal.

Mental health and our word choices

The first writing reviews, started by psychologists in the 1960s, found that patients diagnosed with depression used the pronouns “I” and “me” more often in their text. They also used more words with negative effects, as well as more words related to death. Not surprisingly, all of this early research was confirmed by the most recent algorithmic analysis of mood by the hedonometer and other computer systems.

The University of Vermont’s Danforth, in partnership with Harvard psychologist Andrew Reece, recently analyzed Twitter posts from trial participants who had been formally diagnosed with depression or post-traumatic stress disorder. Using messages published before their diagnosis, signs of depression began to appear up to nine months earlier, the researchers said.

Facebook, often the bane of those who worry about technological overtaking and the erasure of our privacy, has an algorithm to detect posters that appear to be at risk of suicide. In this case, Facebook provides human experts who review the cases and send users prompts or support numbers if they see fit.

Stevie Chancellor, an expert in human-centered informatics at Northwestern University in Chicago, believes that sentiment analysis, despite its many drawbacks, could be useful for clinics, for example, when triage of a new patient.

Measuring our moods is part of today’s consumer culture

Obviously, the business world does not have time to appropriate the tool for its own exploitation. Sentiment analysis is now widely used by businesses, but since many don’t want to admit they are using it, there is no way to know exactly how many are using it today.

Bing Liu of the University of Illinois, who was one of the pioneers of sentiment analysis, said, “Everyone does it: Microsoft, Google, Amazon, everyone,” adding: ” Some of them have multiple research groups.

Today, a large number of commercial and academic sentiment analysis software is widely available. Naturally, one of the data mines that businesses use is social media for the purpose of figuring out what their customers are saying about them and their products.

Naturally, no way to keep tabs on employees is missed by large employers, like IBM, which has developed a program called “Social Pulse” to monitor the company’s intranet to determine what employees can get. to complain.

For privacy reasons, the program only reviewed posts shared with the entire company.

Despite this stipulation, such practices worried Danforth, who said: “My concern would be that employee privacy does not measure up to the bottom line of the business. It is an ethically sketchy thing to do.

No one, however, thinks sentiment analysis using algorithms to gauge our moods is going to go away anytime soon.

However, all companies, mental health professionals and others who employ it, should keep in mind that the human mind is limitless in its complexity.

Understanding humans is a never-ending quest. As Liu says, “We don’t even understand what understanding is.”

About Roberto Frank

Check Also

Top 25 transfers for the 2022-23 season from CBS Sports

A standard 3,616 Division I players have entered the transfer portal for the …

Leave a Reply

Your email address will not be published.