Artificial intelligence and Big Data get lots of ink (and electrons) these days about their awesome promise for the future. But here are two things of real and immediate importance that AI could do today to improve people’s lives and strengthen our economy.
Big Data is a Big Deal because we are being already overwhelmed with information, and it’s going to get a lot more overwhelming very, very soon. The research firm IDC estimates that within two years we will have be awash in an astonishing 40 zettabytes of data (a zettabyte is one sextillion bytes), or 50 times the data that existed just in 2010. This will be the equivalent of 5,200 gigabytes of data for every man, woman, and child on the planet.
Now, various agencies of the federal government manage massive stores of data in order to serve hundreds of millions of Americans. How could AI enhance those agencies’ ability to do their job more effectively? Let’s look at two examples: the U.S. Department of Veterans Affairs and the U.S.
Patent and Trademark Office (USPTO).
The Department of Veterans Affairs has at least 21 million active files, representing each veteran who has ever served our nation. Somewhere in those files lie clues that could help us identify the 22 veterans each day -- almost one every hour -- who are going to commit suicide. But we don’t have the human resources to sift through those files proactively for the key indicators associated with an increased likelihood of suicidal action. Heck, with months’-long waits to even see a doctor, VA staff can barely keep up with the vets coming in for simple check-ups.
But AI expert systems, designed by trained psychologists knowledgeable about suicidal triggers, could do it. They could identify veterans most at risk, and design interventions that could save hundreds, and maybe thousands, of lives.
Now look at the Patent Office. The USPTO has some 14 million files representing the active store of patents issued by the government for novel, non-obvious, and useful inventions. These Patent Office files have been called the greatest storehouse of technological knowledge on the planet. And that's exactly what it is.
Every year, the Patent Office receives more than 600,000 applications for patents, each of which has an average of 15 claims and 75,000 characters. Each of these applications must be individually vetted for novelty, non-obviousness, and utility -- which among other things, means checking them against the 14 million patent files in the USPTO database.
Even though the Patent Office employs 8,500 highly-skilled patent examiners, each of whom has significant expertise in multiple technical areas, it’s a gigantic if not borderline-impossible task. And it’s one that each examiner only has a few hours and very limited resources with which to do it. And because of these limitations, many experts say there's been a decline in patent quality.
How could AI help the Patent Office screen out those applications undeserving of a patent, and ensure only that genuinely worthy inventions receive a patent? For one thing, an AI system could pre-screen and pre-search all new applications looking for direct similarities in their claims to those of already-existing patents, which would therefore spot those that seem to be not truly original. Examiners would of course need to follow up by applying their own human judgment to the evaluation, but AI would certainly save time and improve the examination process.
One could even imagine an AI system that looked for common markers of highly-valuable patents -- perhaps citation frequency, percentage of non-patent prior art cited, or other indicators. In any event, an automated artificial intelligence system could be a boon to the patent examination process and help boost the quality of issued patents.
In point of fact, the Patent Office did sponsor a “bake-off” of just these sorts of AI prescreening and searching capabilities a year ago. Some 22 companies participated, and the contest was won by teams from both the giant consulting firm Deloitte, and from a very promising young AI software startup called Artificial Intelligence Patents (AIP), associated with Duke University in Durham, North Carolina.
The AI bake-off was deemed a success. But somehow in the transition to a new director and administration at the Patent Office this year, progress has stalled. The effort should be revived.
We all have a stake in improving the work of the Department of Veterans Affairs and the U.S. Patent Office. The VA looks after the men and women who risked everything to look after our nation’s safety, and we owe them a lot more support than they are receiving today.
As for the Patent Office, it’s the incentive engine of America’s continued technological greatness, and the key to our future prosperity and our nation’s global competitiveness.
Yes, telling Alexa to turn on our lights or play us a song is great. But the promise of AI is much greater -- and far more important -- than that. Let’s not waste the opportunity.
OUR PROJECT IS BAIM – Business Artificial Intelligence Machine
To Know More About AI we are making please visit-https://www.lvugke.com/
Or watch the video- https://youtu.be/Z5vxRC8dMvs