Dark data is operational data that is not currently being used or analysed, but could potentially provide knowledge of crime rings lurking on the web to creating economic opportunity for companies. That’s what Lattice Data is doing – unlocking this data and turning it into “structured, understandable” material to help governments, law enforcers and companies understand critical real-world problems. Using a statistical inference programme known as Deep Dive, Lattice Data is creating products for industries and government, allowing them to solve data cleansing, extraction, and integration problems jointly.
How they’re disrupting
Unlike traditional machine learning, Deep Dive requires no laborious manual annotations. Instead, it uses existing structured data to solve problems. The aim is not only to match human-level quality of data analysis, but to do so on an unprecedented scale.
52% of data is ‘dark’, holding information that is not available in any other format. An estimated 21 million victims are trapped in modern-day slavery. It’s big business, earning traffickers roughly $150 billion a year. But that’s only one market. Lattice Data is also being used to construct gene-disease relationships from biomedical literature.
Investments and future
Lattice secured an undisclosed amount in its seed capital funding round. The company’s Deep Dive technology was incubated for over six years with $20 million of DARPA funding, which was used to rapidly make sense out of unstructured data. Lattice has since built a suite of proprietary software on top of this to advance the technology further. Since the applications for Disrupt 100 judging closed Lattice was acquired by Apple for an undisclosed amount.