Emotions.Tech is a new organisation, formed in 2017 by three experts with 35+ years of experience in total in Natural Language Processing, Machine Learning, Software Development and Marketing. See our press release
Our goal is to help make the Internet more emotionally transparent. We feel that anger is over represented on the Internet. Driven by the phenomenom of filter bubbles, people's angry views are easily reinforced and amplified. If people were aware in advance of the emotional content of a link or post or comment, they could better monitor and manage their own emotional response to what they are reading.
Ilia Zaitsev, CEO
Ilia studied and taught computational linguistics at Saint-Petersburg State University of Culture and Arts, where he earned his PhD. He now uses machine learning techniques to analyse language, developing models to extract the emotion and sentiment from text. He prototypes and investigates and generally just makes things work.
Paul Tero, CTO
Paul studied computer science at UC Berkeley and Evolutionary and Adaptive Systems at the University of Sussex. He is a web developer turned data scientist, who understands the maths behind the models.
Michiel Maandag, CMO
Back in 2004 Michiel was working as a business developer at Nokia where he led the project that created a working prototype of what we would call today Siri. These were the early days of running AI on a phone. The project was never released to mass consumers, but the passion for AI stayed. His career in branding and marketing started at Nokia in 2006 where he led brands and managed teams globally. Through his own company he has served many companies since 2012. In 2015 his book The Only Book You Will Ever Need on Branding became the 'non-fiction book of the month' in the UK and a top 10 marketing bestseller in the Netherlands.
Cliff Crosbie, CXO
Cliff Crosbie is a retail and marketing professional who has been privileged to work with some of the best brands in the world over the last 30 years. He ran Apple's global premium reseller and shop in shop program, where the focus was on delivering an exceptional experience for customers in over 90 countries Previously Cliff was Japan Country Sales Manager for IKEA, Global VP for Nokia's Retail and Customer Marketing, and Head of Nike's European Franchise program. He was also the first GM of Niketown London, perhaps one of the first true omnichannel experiences. Cliff is focused on improving the customer experience in all forms of business through the use of cutting edge technology - the convergence of online and offline retail is where it's all happening today.
In March 2017 we were approved to join the Nvidia Inception program, a business incubator which helps startups using machine learning on Nvidia graphics cards.
We are working with the independent no-tracking search engine Mojeek to add emotions to their search results. Mojeek has crawled well over a billion pages, making it one of the largest English search indices in the world.
Machine learning is an old branch of computer science which has recently become very popular, mainly because the speed of modern computing makes it practical. Instead of humans writing algorithms and equations to solve problems, the machine analyses mountains of data and learns the equations itself. The equations it learns take the form of a neural network, also known as a model.
The field of Natural Language Processing uses algorithms to study natural languages like English. For example sentiment analysis algorithms take a sentence like "I feel great" and programmatically estimate whether this has a positive or negative sentiment. Traditionally, this is done by counting up positive and negative keywords, which works for "I feel great" but not "I am over the moon".
At Emotions.Tech, we are successfully applying machine learning techniques to NLP tasks to overcome this problem. So far we have developed two models, one for emotion analysis from the reader's point of view, the other for sentiment analysis from the writer's point of view.
Currently very few companies provide emotion analysis, and none of them draw the subtle but important distinction between reader and writer. For example, advertising companies are not really concerned with how the person who wrote the ad was feeling at the time, they are much more interested in how the reader reacts to the ad. A good example of this is JCPenney's 2013 ad campaign for a kettle which resembled Hitler. JCPenney were very worried when this emerged, but the general response on the internet (and from our emotional model) was amusement.
Our Angry Blocker extension is our first direct attempt to fulfil our goal. The emotional web allows people to analyse, search or discover the Internet by emotion. We are also interested in brand safety via emotional targeting. Our emotion analysis model provides the core functionality for all these products. See our products page for more information.