Founded in 2008 after more than a decade of research at Stanford, Silicon Valley-based startup, Ayasdi, is on a mission to reinvent the methods by which we transform Big Data into actionable knowledge. Essentially, Ayasdi aims to automate the insight discovery process, allowing end users to find valuable intelligence within massive datasets almost instantaneously. Backed by the Defense Advanced Research Projects Agency (DARPA) and the National Science Foundation, the startup’s novel synthesis of machine learning and data analysis technologies has not gone unnoticed by investors.
Today, the startup announced that it has raised $10 million in series A financing, led by Khosla Ventures and Floodgate. The new round brings the startup’s total funding to $13.25 million, which includes contributions both from its current investors as well as angel investors like Michael Ovitz, Steve Blank and The Data Collective’s Matt Ocko, to name a few.
The investment comes at a time when venture capital is beginning to shift from consumer businesses to the enterprise, as investors look to capitalize on the rise of the big data app and the growing demand for a new data infrastructure.
In fact, Research firm IDC recently predicted that the Big Data market is poised for exponential growth over the next few years, with total revenue projected to reach $24 billion by 2016. What’s more, following Derrick Harris’ logic, if one includes analytics software (arguably an essential segment of the Big Data market) in that definition, then the total is actually upwards of $75 billion.
Yet, in spite of this enormous new market opportunity, the emergence of Big Data doesn’t necessarily directly translate into a net positive for business, humankind, etc. In other words, people tend to conflate information and data, such that we end up believing that more data inherently means (proportionately) more information, which in turn means (proportionately) more intelligence, insight and value. However, as Dr. Michael Wu points out, it’s quite literally the opposite: “The more data you have, the less information you gain as a proportion of the data.”
The true value of Big Data is derived from the insights hidden within, yet, while all insights are information, not all information produces insight. Big Data is composed of enormous amounts of unstructured data, a wide array of data types and media, but the amount of insight that can be extracted from that data is proportionately tiny.
Big Data continues to grow, yet, while governments, businesses and scientists have spent years (and millions of dollars) attempting to address the world’s biggest problems by analyzing Big Data, progress has been incremental. Although Big Data tools have improved over time, Ayasdi is of the mindset that they are still failing to yield the kind of breakthrough insights that lead to true innovation.
The Ayasdi co-founders attribute this to the prevailing reliance among data scientists on old models — finding insights by asking questions and writing queries. The problem with this is that queries are inherently based on human assumptions and biases, and, in turn, query results tend to only reveal slices of data, rather than providing visibility into the relationships between similar groups of data. This method of discovering insight in Big Data tends to rely heavily on iterative guesswork and chance, and thus takes time to produce real results.
To address this problem, Ayasdi is today officially launching its cloud-based insight discovery platform, which aims to deliver insight derived from massive datasets quickly, without relying on queries. The machine learning platform combines computer science with a branch of mathematics called “Topological Data Analysis,” which allows Ayasdi to visualize entire datasets at once.
The startup’s platform uses hundreds of machine learning algorithms to explore these complex datasets — the goal being the automatic discovery of insights that could not be discovered through ad hoc or query-based methods. The platform, which is designed for domain experts, data scientists and researchers, requires no coding or modeling and offers the kind of scalability that more demanding processing requires. It also is built to complement other Big Data solutions companies might already be using and is able to work with datasets of any size or type, the co-founders tell us.
The startup is currently working with enterprise customers in financial services, life sciences, oil and gas and the public sector, with these companies employing its platform to help discover new drugs, for example, improve cancer therapy by discovering new insights from an 11-year old breast cancer dataset that included new sub-populations of breast cancer survivors, for example. It’s also being used to explore new energy sources, identifying patterns that can lead to more accurate drilling, predict fraud and help prevent terrorist attacks.
“The answers to today’s most important scientific, business and social problems lie in data,” saus Ayasdi CEO Gurjeet Singh. “The biggest challenge in Big Data today is asking the right questions of data, so the real opportunity in Big Data lies in the automation of insight discovery — regardless of the complexity of that data — without requiring users to ask questions. The goal is for Ayasdi to provide users with answers to questions that they didn’t know to ask in the first place.”