NEC's RAPID machine learning supports human decisions with artificial intelligence, improving work efficiency!
- TomonagaFor example, when hiring new graduates, in the past a professional would have looked at the application forms and aptitude tests of applicants, and decided who to interview and who to turn down. We replace this with a computer. RAPID machine learning is trained with a model for the people a given company wants to hire, and used to compare data such as the application forms and aptitude tests of new graduates, to support decision-making with regard to whether or not to move the next step.
- FujishigeThe companies hiring also incur labor costs through document screening and interviews, so they would like to reduce these by as much as they can. Other benefits are the potential to secure gifted personnel at an early stage, and prevent their omission from consideration.
- MitaI see.
- TomonagaAnother application example where machine learning shines is the need to forecast performance a few years down the line when a new recruit joins. Companies would like to simulate at the time of employment whether a new graduate will be successful after a few years in a given department if assigned to it.
- FujishigeWe believe this could be handled by creating a model that ties together the results of application forms and aptitude tests for new graduates hired in a certain year, with their performance a few years after joining the company. Comparing this model with the results of application forms and aptitude tests for incoming new graduates would make it possible to forecast which department they would likely do well in.
- MitaWow, being able to predict someone's future sounds exciting.
- TomonagaWith regard to human resource matching, we are still at the stage where we are exploring what can be done, and looking into the requirements we can meet, together with company personnel management officers and other collaborators. However, we are convinced that RAPID machine learning has a huge amount of potential for utilization in a range of fields.
- MitaBy the way, you said earlier that RAPID machine learning can be run faster and with less memory than other deep learning software, but what benefits does this create when used at companies?
- TomonagaThat's a good question. Most machine learning tools available from competitors are cloud-type solutions. In other words, to analyze data such as images or text, that data must first be uploaded to the cloud. That means using these tools is more difficult for companies that handle confidential information, such as surveillance camera image data and corporate personnel data, for example.
- FujishigeAlthough safety is increasing, many companies are concerned about security when it comes to putting security camera footage or personal information for human resources matching in the cloud.
- MitaI see.
- TomonagaDue to this, a large number of companies are looking for software they can incorporate into their internal system, and because RAPID machine learning is very memory and resource efficient when running, it can be introduced to internal systems easily. In summary, it can also be used by customers who aren't able to take their data outside the office.
- MitaThat's a big advantage!
- FujishigeAlso, when trends for the data to be analyzed change quite often, data learning must be implemented frequently. RAPID machine learning can be done swiftly without the need to teach data features, so it is suited to cases such as this.
- MitaIndeed! In the near future, it is likely our lives will change significantly due to the spread of software and systems based on artificial intelligence technology like this. I'm looking forward to it! Mr. Tomonaga and Ms. Fujishige, thank you for taking time out of your busy day to share your valuable insights with us!
- Tomonaga, FujishigeIt was a pleasure answering your questions. Thank you!
In this installment, we heard about software utilizing the artificial intelligence technology that is currently a hot topic. I marvel at the fact decisions made by people will now be able to be made by computers. It won't be long before devices equipped with this artificial intelligence start being used all over the place, and this could be the advent of a society that offers convenience beyond our wildest dreams. I'll definitely be keeping my eye on future developments in technology related to artificial intelligence. See you in the next installment of "MiTA TV"!
(Published September 15, 2015)