Luminoso, a Cambridge, MA-based text analytics company announced $6.5M in Series A funding today, led by Acadia Woods with Digital Garage also participating. That brings their total funds raised to $8M.
Luminoso was born out of research at the MIT Media Labs and spun off in 2010. After getting their seed funding in 2012, they have continued to develop the product, which provides text analytics in the cloud — and this money should give them a big lift to continue building the company and hire more personnel.
What makes it unique among its competition is that they have removed the hardest part of the text analytics from the equation by automatically building a taxonomy on the fly. According to CEO and co-founder Catherine Havasi, they built data modeling and machine learning so it’s all automated, a pretty neat trick if they can pull it off.
That means according to Havasi, the system builds its own taxonomy of words, and even when slang enters the lexicon, Luminoso is capable of recognizing it as a new word, and inferring its meaning from contextual clues of users.
“People are creative about the way they talk about a product,” Havasi said. “We wanted to give computers intuition to understand creative language. At the Media Lab, we built a set of data that acts as a model of how people think about the world,” she explained
That resulted in developing modern machine learning, but they took it a step further and made it cloud-based to make it easier for customers to tap into.
So far, she says, the biggest use case has been around market research and understanding streams of text as they come in to gain a better understanding of what’s happening in the market. While she said, it’s applicable to a lot of usage scenarios, as a startup it was important to focus on a strong use case and build the company from that.
The goal of course is to help people make sense of large amounts of unstructured data and so far, at least the focus has been on understanding customer interactions. Havasi says as a proof of concept they have been working with Sony compiling information from social media about people’s reaction to the World Cup as it happens.
As a New York Times article on using this technology for the Germany-US match last week pointed out, it’s more than simple sentiment analysis, as in people are happy or not. “Its technology also includes a knowledge base, which is built off a crowdsourced project at the Massachusetts Institute of Technology that began in 1999 and is called the Open Mind Common Sense Project,” according to the article.
Havasi says she sees other text analytics companies like Lexalytics and Clarabridge as the company’s primary competitors, but she said the major difference between her company and the competitors is they use people to build the taxonomies and Luminoso is fully automated (and continues to learn even as the taxonomy changes and evolves).
Havasi said, this takes the human element out of the equation as far as taxonomy creation and analysis is concerned, but not altogether. She believes when you put a human and a computer together with technology like this, it will give people better insight into their customers (or however they choose to use the Luminoso technology), while automating the hard parts that take people a long to time to build and process.