A huge part of human communication relies on empathy and reading non-verbal signals. We share joy and sadness with others, even if we don’t intend to. That’s why reading emotions has always be a barrier in human-machine communication. But today we are getting closer and closer to solving the mystery.
Internet of Things, or how the smart fridge will affect your job. Download Free Ebook
Between EEG and emotions
For a long time, we couldn’t translate brain waves into emotions. How to interpret EEG? Finally, Wei-Long Zheng and his team undertook the problem. Researchers attached electrodes to subjects’ heads to analyzed electric signals, recorded subjects’ facial expressions and asked them to describe their sentiments towards the movies they were shown. Correlating all the data, Zheng found patterns describing what type of brain waves is responsible for what emotions.
Although the subjects can’t be entirely representative (we don’t know for example if the patterns identified would apply to older people), but it’s a big step ahead.
An algorithm that detects sarcasm (better than humans)
People tend to overestimate their ability do express emotions openly. What seems obvious to you in your statement, might be obscure or vague to your interlocutor.The internet only escalates the problem. In the 2005 research, participants identified sarcasm accurately only in 56% of the cases!
A significant problem for scientists to address: how to find a relation between words and emotions whey evoke? The most effective answer was given by David Bamman and Noah A. Smith, who believed that not only the statement itself must be examined, but emotion-detecting machines must also take the context into consideration.
Bamman and Smith started from Twitter, searching for tweets with hashtags “#sarcasm” and “#sarcastic”. It turned out that author’s profile and target audience were the most important factors shaping the meaning.
Their research is now used among others by Department of Defense. It helps fish out the potentially dangerous remarks from offensive jokes.
Commercial use: Affectiva
Affectiva is a tool that tracks users’ emotions in real time with video recorders built in various devices (from computers to smartphones, to Google Glass). The system interprets mimics into emotions, so it identifies what products or pieces of content engage, moves or entertains us most.
Such solution is applied in many areas: to optimize and personalize content in real-time in response to particular users’ reactions. In the film industry, Affectiva is used to pick the best movie moments to compose trailers.
Manipulation
But with great power comes great responsibility. Will brands and media gain too much room to manipulate our emotions? In the famous (or rather infamous) experiment two years ago, unaware Facebook users were split into two groups: one was given more positive content in their feed, the other one – the opposite. The portal didn’t interfere with the content, but slightly modified the mechanism of displaying and selecting posts. The project involved only curating, not creating or modifying posts.
The decision to run tests on Facebook users without asking them for permission divided the public opinion. One side defended Zuckerberg’s platform, saying that each day we’re subjected to similar manipulations and experiments, like A/B testing or personalization techniques. Whoever wants transparency from the Internet shows an enormous dose of naivety and can’t blame reality that it doesn’t meet such improbable standards.
The opponents claimed that it was about trust – we share with Facebook our intimate moments, our relationships and emotions, so nothing can justify making use of that data cynically, even if it’s about the science.
For sure, the new solutions will change the marketing, too. We will know more about how our customers respond to our messages and products, and how they like given stages of the shopping process. We will learn what elements of our website do they like, and what doesn’t raise such enthusiasm, what they enjoy and what irritates them. Think how will it enhance Marketing Automation: specialists will be able to customize the content and the message according not only to user’s profile and history but also to how she feels right now.
How do you like that? Is it delightful or rather scary? Let us know in comments!
Follow
