The analytics outfit that helped uncover North Korea’s biological weapon secret
Fri 18 Jan 2019
By delving into deep web data, Wales-based AMPLYFI can derive links no human could ever identify. James Orme goes on a data voyage with VP of engineering, Joe Mathews
According to industry analysts, in a matter of months something unprecedented will happen in the land of dragons: one of its fledgeling tech startups will achieve unicorn status. Despite a surge in tech startups calling Wales their home in recent years, intelligent analytics firm AMPLYFI’s impending coronation has led some commentators to confidently proclaim the dawn of a new digital era in Wales.
Ever since AMPLYFI moved to Cardiff in 2015, it has comfortably completed one funding round after another, thanks in the main to its flagship analytics tool DataVoyant. Although currently hiring engineers, a cursory glance at the AMPLYFI LinkedIn reveals an employee count of 35. It’s quite remarkable that within five years, this state-of-art shoebox is expected to hit £50 million in turnover.
Using artificial intelligence DataVoyant harvests the internet to find, harvest and analyse data, which businesses can then use to spot and react to trends and disruption. While it sounds like a page straight out of the analytics playbook, what distinguishes DataVoyant from its competitors is its unique use of both deep web and surface web data.
“No other analytical tool does this,” says Joe.
The deep web is often confused with the dark web, the notorious underbelly of the internet accessible only by specialist web browsers such as Tor or I2p. But the deep web is not a meeting ground for nefarious actors planning cybercrime or a testing ground for lofty libertarian visions of capitalism. It simply refers to the parts of the internet cut off from standard search engines – where you find mundane things like academic papers, financial records and legal documents; not counterfeit passports or Amsterdam’s top-shelf THC.
Stepping into the unknown
AMPLYFI’s revelation was that this vast pool of data represented a largely untapped source of intelligence and insights, that could help businesses and governments spot altogether new trends.
But there was a good reason no other firm had created a tool to translate this wellspring of information into insight: its sheer size and depth.
“Using deeper insights than they have ever had access to, our clients can make informed decisions about any strategic focus”
“It expands exponentially beyond the limited results presented by a typical surface web search engine such as Google or Bing,” Joe explains.
“Its size defies quantification, but the deep web is estimated to be around 400-500 times bigger than the surface web. Harnessing this endless wealth of potential information and translating it into useful business insights is an intricate and complex process. With AMPLYFI’s DataVoyant technology, we’ve found the way to do it.”
The way to do it, unsurprisingly, involves using machine learning to process data at speed and at scale. DataVoyant launched after just one year of research and development, and performed better than AMPLYFI could have ever hoped.
“DataVoyant can look at tens of thousands of datasets in the time that a human can review just a handful. This is a major shift for business analytics: we’re helping the analysts get on with analysing, enabling those teams to be more productive, effective and impactful.”
But it would not do justice to DataVoyant to pin its power on clever algorithms and the unconventional data source it feeds off. The software uses panoramic big data visualisation to create a platform that is intuitively accessible for conventional IT teams, and one that is also able to understand and interpret most modern languages.
“DataVoyant can spot and derive links in data that no human could hope to identify. These are the ‘unknown unknowns’,” Joe says.
The platform now boasts thousands of users worldwide, a number Joe hopes to take into the hundreds of thousands or even millions. It’s not a radical ambition, given how valuable the technology is to virtually all industries. There is not one single industry that couldn’t benefit from knowing what trends are approaching, and the best ways to respond to them.
“The most obvious thing to conclude from all of this is that if DataVoyant is good enough for the US military – it could likely do wonders for your business”
Joe boasts that DataVoyant is already influencing strategic business decisions at the highest levels across a whole range of industries, from finance and banking to energy and pharmaceuticals.
AMPLYFI is also making inroads into the public sector. A project with the Singaporean Government is helping them predict the skills and roles that will be needed by its workforce up to 50 years in the future. Insights Joe says could change their approach to education, skills and development for ‘decades to come’.
“Using deeper insights than they have ever had access to, our clients can make informed decisions about any strategic focus, whether that’s their R&D roadmaps, acquisitions, new markets or another critical topic,” he says.
The applications for a tool of this power are really limited to a firm’s imagination. At the end of last year, seemingly for a bit of fun Christmas marketing, AMPLYFI used DataVoyant to predict the next crop of technology megatrends – its AI-powered answer to the ubiquitous technology predictions that annually circulate the industry.
In finding, harvesting, and reading over 1,000,000 open-source documents and identifying over 2000 themes, it software predicted AI, IoT and DLT technologies to have the greatest impact in the coming years. Which at least reassures Techerati that we’re on the right track. Fully autonomous cars? According to DataVoyant best not to get your hopes up – they are at least 10 years out.
But it is AMPLYFI’s collaboration with Harvard University that brings its power into sharp relief. In its work with the Ivy Leaguer, DataVoyant has proven itself as a tool for state defence.
Harvard used DataVoyant to mine over 840,000 websites as part of research into the potential biological weapons capability of North Korea. It found 40,000 websites that contained broad biological references, of which 23,000 were revealed to have associations with North Korea, and 170 that established a direct link to regime organisations and institutions.
DataVoyant also found incriminating search activity, namely a marked increase in searches for “antibiotic resistance,” “microbial dark matter,” “cas protein” suggesting a growing interest in advanced gene and germ research.
Last month, AMPLYFI’s work was cited in military analysis issued by the Middlebury Institute of International Studies. DataVoyant’s analysis forms part of a constellation of evidence suggesting Pyongyang is developing an “advanced, underestimated and highly lethal” biological weapons program.
The results even led to Vernon Gibson, former Ministry of Defence Chief Scientific Adviser showering praise on Cardiff’s king web crawler.
The most obvious thing to conclude from all of this is that if DataVoyant is good enough for the US military – it could likely do wonders for your business. In order to meet the inevitable future demand for DataVoyant, we can expect AMPLYFI’s Cardiff HQ to expand significantly in the coming years.
“Our biggest challenge has been scaling the company to cater for the huge market interest we have received. In just three years, we’ve grown the business to more than 30 employees, we’re already working with more than 50 customers and have potential projects for perhaps 85 more. They’re fantastic challenges to have, and long may they continue,” says Joe.
Tags:Big Data dark web data analytics deep web feature
Big Data Fri 18 Jan 2019How Marks Spencers leverages data science to reduce fo...
Exploring the security benefits of hybrid cloud
Read More >>
Why we need to automate automation
Read More >>
Collaboration with young generation key to virtual bank success, says ZhongAn...
Read More >>
Before you pay your ransomware read this
Read More >>
Are you prepared for the enterprise 2G/3G shutdown?
Read More >>