Context is critical. As what was once mere data evolves into actionable intelligence, the context that binds that data becomes ever more essential.
Consider the word “java.” With no context around those four letters, you might not understand the reference or make any sort of connection. But if you add just one word to “java,” such as “development,” “island,” or “coffee,” the reference changes completely—and that’s with just a single word of context.
This is the type of active context and connection that the Brainspace engine provides. “Context is a very important part of what we do. When we analyze documents, we take the context into consideration,” says Ravi Sathyanna, vice president of technology and product management at Brainspace. “The question is: when you have unstructured data, how do you actually go about analyzing it? We do concept searching. We are able to analyze tens of millions of documents and build relationships between concepts in all of those documents.”
Brainspace can create enhanced semantic queries from a single term, a sentence, a paragraph, or even an entire document. As the Brainspace engine parses and analyzes structured or unstructured data, it derives concepts and context from that data. “Everything is dynamically derived and defined from the data customers give us,” Sathyanna says. “That could be from documents within the enterprise content management systems like SharePoint, or from e-mails or news articles. We don’t start the process with any known relationships.” Given the dynamic nature of the Brainspace engine’s machine-learning capabilities, customers can feed any type of unstructured data into the system.
“The core Brainspace platform is an unsupervised machine-learning environment that learns dynamically without the use of any pre-built lexicon, ontology, or thesaurus,” says Brainspace CEO Dave Copps. “Our platform can ingest unstructured data on a massive scale. We have built queries today that analyze and connect the concepts inside of hundreds of millions of documents that allow people to explore their big data in ways that have never been possible before.”
Once the context has been defined, the Brainspace analysis also provides weighted relevancy for specific documents. “We provide relevancy—and, equally important, transparency—as part of the concept search. You can execute a query, evaluate the most related concepts, and optionally balance the weights to impact your search,” Sathyanna explains. That process helps human analysts determine the relative significance of particular documents during search and analysis.
The scale and speed at which Brainspace can ingest documents significantly differentiate it from its competitors, Copps notes. “Ingesting and building the initial Brainspace from a million documents, for example, takes about 30 minutes with no human intervention,” he says. The Brainspace engine can also learn in multiple languages, he adds: “We’re optimizing for 20 major languages, including Mandarin, Kanji, Korean, and Farsi. Our platform not only learns natively in those languages, but automatically identifies phrases, increasing our ability to extract meaning in even the double-byte languages.”
Consider one example of how the technology works. LexisNexis, a well-known provider of legal, government, and business documents, used the Brainspace engine to ingest and learn from every patent ever issued in the United States and Europe as well as from millions of journal articles—a total of more than 350 million documents.
Visual analytics is another unique facet of the Brainspace process. One example of this capability is showcased via document clustering. Sathyanna, who refers to visual analytics as “unsupervised learning,” describes the process this way: “You give us any number of documents. We analyze them and group them into clusters. Then you’ll have a visual representation of these clusters. That’s a visual representation of the entire document population in a grouping with labels on them. It allows you to navigate through the data set much like you can navigate the world via Google Earth.”
The combination of search capabilities and machine learning is critical to the visual data analytics that Brainspace delivers. The solution brings a large amount of data together in a visual representation, making it easier to understand. “What I want to do is enable people to have a conversation with their data,” Sathyanna says.
The user experience makes Brainspace that much more accessible, Copps adds. “It’s one thing to build great machine-learning technology. It’s a whole other thing to be able to present that in a way that enables people to actively engage with it,” he says. “The key to understanding the enterprise lies in our ability to understand unstructured data. That’s where the stories and ideas that drive an organization live. Through a meaningful user experience, we are creating the bridge between machine learning and human curation that enables enterprises to finally reach that understanding. The user experience is simplified to the point where training can happen in minutes, as opposed to hours.”
The Brainspace environment presents data visualizations with all documents clustered in the center of a wheel. As users click on a data cluster, other sub-clusters appear, taking users deeper into the details of their discovery. “It’s a unique environment and user experience that makes it possible to visually navigate large data sets,” Copps says. “We’re expressing machine learning in ways from which virtually anyone can extract value. You don’t have to be a data scientist to use our products.”
Building inference and related concepts into queries can be particularly helpful for document searches related to social-media outlets. “We provide conceptual search functions that automatically bring inference to the surface as it relates to the concept to build into queries,” Copps says. “It’s like stepping into the middle of a multi-dimensional sphere and being surrounded by words and phrases that are organized by their relevant distance to each other.”
This type of data analytics is particularly useful in research-heavy applications such as legal e-discovery, fraud detection investigations within financial-services organizations, and compliance or governance issues. “Our engine can automatically look at the documents, analyze them, and put them into the appropriate buckets, considered within the policies of enterprise,” Sathyanna says.
Supporting litigation is where Brainspace really got its start. “We cut our teeth in e-discovery and have become the go-to analytics technology for large litigations, like when things blew up at Volkswagen,” Copps says, referring to the German automaker’s ongoing diesel-emissions scandal. “Our platform takes companies from analyzing millions of documents down to just the few documents that matter more quickly than any product on the market today. Seventy-five percent of the costs associated with discovery are tied to human review, and people are expensive. So by radically reducing the number of relevant documents early in the process, we are dramatically lowering the cost of review.”
Such deep investigative activity is integral to Brainspace’s Discovery 5 product. “Discovery 5 is intended for data analysts and investigators, and is used extensively in investigations and e-discovery,” Sathyanna says. Another product based on the same core engine is Brainspace for Enterprise, which is re-inventing enterprise knowledge management. As users produce, collect, and share knowledge, they form a unique collective intelligence—a “Brainspace”—that everyone can use to connect more meaningfully with relevant knowledge and peers. It’s built specifically for the broader swath of knowledge workers inside the enterprise.
“Discovery 5 serves data scientists,” Copps says. “It creates a visual-analytics environment that allows them to explore unstructured data in ways that previously were only possible in more structured data environments. Brainspace for Enterprise, on the other hand, helps knowledge workers curate, collaborate, and discover information and people inside a dynamic learning environment. Both products sit on top of the core Brainspace technology.”
Discovery 5 and Brainspace for Enterprise provide what Brainspace calls “augmented intelligence,” an evolution of artificial intelligence, or AI. Brainspace incorporates machine learning to supplement and support human analysis, Sathyanna explains. “While the system can learn without human intervention, we are also augmenting the users’ decision-making process and capabilities by providing deep insights which are otherwise hidden or inaccessible.”
This level of intelligence represents the current evolution in machine learning and AI. “Augmented human intelligence is the next step in AI,” Copps says. “We have reached the point where the synthesis of machine learning and human curation has the potential to completely reshape data analytics in the enterprise.”
Copps describes it as bringing the best of both worlds—machine and human capabilities—together. “A machine’s ability to ingest, connect, and recall information goes far beyond what is possible for humans,” he says. On the other hand, a person’s ability to use information to reason, judge, and strategize far surpasses the capabilities of machines today, he adds: “By combining these abilities inside of an augmented intelligence environment, we are able to accelerate productivity beyond what is possible with other, more traditional tools.”
In other words, the intelligence provided by Brainspace’s analysis helps amplify the level of human intelligence. It can help human analysts rapidly draw conclusions that might have been impossible to reach before—or, at the very least, to do so much faster.
For more information on Brainspace, visit www.brainspace.com.