Jeff Bertolucci of Information Week recently asked, “So what’s a smart machine? As a computing model designed to analyze growing volumes of unstructured data, including video, images, and human language, a smart machine, or cognitive computing system, uses artificial intelligence and machine learning algorithms to ‘sense, predict, infer and, in some ways, think,’ according to IBM. In a recent report, Cool Vendors in Smart Machines, 2014, Gartner named three well known examples of smart machines, including IBM’s Watson, Google Now, and Apple’s Siri, and predicted the technology will gain wider acceptance this decade. It also spotlighted a few up-and-coming vendors in the genre, including AlchemyAPI, Digital Reasoning, Highspot, Lumiata, and Narrative Science.” Read more
Posts Tagged ‘Synthesis’
Synthesys, Digital Reasoning’s machine learning platform that ferrets out meaning in unstructured data at scale, is bringing its smarts to compliance use cases for organizations, such as financial institutions. (See this article for more insight into the technology behind the company’s software.)
This week, the vendor delivered Version 3.7 of the Synthesys software, which brings with it the capability to monitor and analyze all email communications in near real-time. That matters to many compliance program use cases, among them insider trading, money laundering and reputation management. “They all go back to finding information inside of communications, like who are the people and organizations mentioned in email, and what is being discussed about them,” says Tim Estes, chairman and CEO. “Synthesys can take essentially millions of emails and winnow them to maybe a hundred that are problems.”
That means fewer things are falsely flagged as issues, there’s less privacy treading into innocent emails, and there’s more return on time for the people charged with protecting customers and enforcing compliance requirements.
Digital Reasoning, developers of the Synthesys platform for discovering the meaning in unstructured data at scale, has on the roadmap exposing to and packaging up for its customers a simplified version of its internal technology for teaching the system new grammatical structures so that it can quickly understand custom or otherwise specific data sets.
The company has quickly added support for new languages such as Arabic, traditional and simplified Chinese, Farsi and Urdu (with more languages on the way) to Synthesys using the tool. The tool gets the software up to speed on each one in just a few weeks by teaching it the grammatical structure and then letting it go off and figure out what the words mean for its work of transforming unstructured (and structured) data into the underlying facts, entities, relationships, and associated terms.
“In the same way we teach it languages you may have a data set that is highly scientific, for example, and this tool essentially makes it easier for our customers to make Synthesys even more accurate for that specific set of data,” says Dave Danielson, VP of marketing.
Are you starting to hear more about patents that relate to the Semantic Web space? There was an interesting discussion by Erik Sherman here on Facebook’s patent for automatic search curation as feeding its semantic search ambitions, for instance.
Generally speaking, in fact, patents are big in the news, with the passage last week by the Senate of the Patent Reform Bill, which has among its goals getting patents issued sooner — but which also is spurring concern, especially in the tech industry, about its impact on patent infringement actions.
Against this backdrop, and perhaps flying a bit more under the radar, was a U.S. patent (No. 7,882,055) granted to Digital Reasoning for its distributed system of intelligent software agents for discovering the meaning in text. Company CEO Tim Estes calls what the vendor has applied to its Synthesys technology a “bottom-up” patent.
Specifically, it covers the mechanism of measurement and the applications of algorithms to develop machine-understandable structures from patterns of symbol usage, the company says, as well as the semantic alignment of those learned structures from unstructured data with pre-existing structured data — a necessary step in creating enterprise-class entity-oriented systems.
So, in plain(er) English, it’s about using algorithms to bootstrap the creation of semantic models from large-scale unstructured data with minimal a priori information – in other words, to let the data speak for itself. It aims at being a fast route to entity-oriented analytics for harvesting critical facts and relationships across a spread of information in documents.