'Verdict' Helps Databases Answer Queries 200x Faster with 99% Accuracy

Databases are getting so big that a problem arises when considering how best to store data and access it. Thankfully, researchers may have a solution that improves the speed and accuracy of existing databases.

There are more things that we depend on databases for than I can list in one article. Data is the foundation of the digital world, and much of the progress that we’ve seen in Industry 4.0 is due to a greater understanding of the transfer, storage, and use of data.

Or, in general terms, you can just call it Big Data.

It is the underpinning of the most advanced sciences and the most convenient tech toys of the digital age. It’s not called ‘Big Data’ for nothing.

To put it into perspective, we have so much data that we’ve invented AI in order to sort and retrieve it. Without these digital librarians, it would take forever for people to sort through all the data that fuels research and commercial ventures. That would be a problem.

Lucky for us, researchers at the University of Michigan have a solution in mind, and it’s called Verdict.

Verdict is a software with a unique purpose: to help existing databases sift through data to retrieve answers. It seems simple, but it’s actually the first entry in an entirely new field of research called ‘database learning‘.

But before we go too far into the solution, let’s try and outline the problem.

The Hazards of Big Data

Let’s take a moment to think about how Big Data got its name.

Currently, there are more than 1 billion gigabytes of stored data in the world. That amount of data is rising rapidly, too rapidly for systems to keep up. And upgrading our processors won’t fix the problem.

Add to that a dependency on data as machine learning software gets more advanced and demanding, and you have the makings of a worldwide data crunch. Everyone from medical researchers to online retailers is making extensive use of Big Data. Without some way to lighten the load, their efforts might be slowed down considerably.

Of course, we may not have to worry about that. Verdict is a software designed to work with existing databases, and it can provide some pretty impressive increases in efficiency.

The Verdict is in

Verdict is an elegant solution for improving existing databases around the world. It is “thin layer” software: simple, adaptable, and small.

You can layer it atop any existing database, for starters, and it has been demonstrated to increase access speeds by as much as 200 times. This is all while maintaining a 99% accuracy rating, according to Barzan Mozafari, the Morris Wellman Faculty Development Assistant Professor of Computer Science and Engineering at the University of Michigan. And if you don’t need more speed, Verdict can cut down power requirements on your database by 200 times.

“Verdict allows databases to deliver answers more than 200 times faster while maintaining 99 percent accuracy. In a research environment, that could mean getting answers in seconds instead of hours or days.”

The software works by changing the way that a database searches for the correct information. It stores queries so that it can create what it calls a ‘query synopsis’. This synopsis develops over time, giving the database a better sense of judgment for each query put to it.

After it has stored so many queries, the software uses them to create a model that will help direct the database to wherever an answer may be found. It takes a page directly from machine learning by letting the machine learn from its past operations.

Basically, Verdict allows databases to gain experience. Will they level up?

But seriously, this kind of technology could help researchers crunch enormous amounts of data in a fraction of the time. That turns a possible data crisis into a potential technology boom.

But how long will we have to wait for this kind of tech to reach the commercial market, and what might we expect from software like this in the future?

Waiting for Verdict

Mozafari says that a commercial version of Verdict might be a few years away. So we may have to wait on that improved digital librarian we’re all hoping for.

But I’d imagine that anything that relies heavily upon a database will benefit greatly from something like Verdict. For example, digital assistants might start to feel like they are reading your mind, or you might start seeing Watson diagnosing patients in an emergency room near you.

Verdict might even offer solutions for self-driving car AI platforms. I, for one, would love to be able to nap on the way to the office.

I, for one, would love to be able to nap on the way to the office.