Hewlett Packard Enterprise (HPE) announced today that it is adding a feature called natural language question answering to the HPE IDOL analytics engine. IDOL takes input in the form of text, images, video, faces, and speech and pulls out trends and insight from that information within an enterprise.
Natural language question answering allows for people to request information, speaking in a normal, non-rigid manner. IDOL is then able to look at data sources from an enterprise’s applications, databases, or documents and puts that information into context based on the user’s question. It is then able to build an index that it queries and pulls data from.
Unlike other search engines, IDOL doesn’t use keywords because keywords are never used in context, says Jeff Veis, HPE Software’s vice president of marketing for big data platforms.
While products like Siri or Amazon Echo have similar capabilities, “This innovation has frozen for enterprise use,” Veis claims. “With enterprises, context is everything because answers need to vary depending on specific questions relating to the same topic.”
IDOL has three engines that are contained in what HPE calls an answer server, and users will get a text-written response after analyzing that question in one of these three engines.
- IDOL Answer Bank is for reference questions that are made to be answered in multiple ways. Using machine learning, it is able to understand the different forms a single question is asked in and can produce an answer.
- IDOL Fact Bank looks for fact-based questions from documents, apps, or emails. For example, if a broker wants to know how much wheat was trading for on a specific day and time, IDOL will be able to find out.
- IDOL Passage Extraction is used to receive a general synopsis that is based on a topic. For example, it can look at when a particular regulation or law was passed and send back the most relevant summary of that topic.
All three of these engines sit on the IDOL server and will look at the question to determine which engine is best suited to answer it.
IDOL has customers in all verticals, HPE claims, but the company highlights those in the financial sector. For example, Morgan Stanley uses it to analyze millions of emails about insider training and trains IDOL on previous patterns of insider trading to get the most relevant information.
IDOL, part of HPE’s Big Data Business portfolio, has been around for eight years and is able to speak in more than 80 languages.