Korea IT Times' interview: BPU's Pioneering AEI Engine and the Real State of AI

BPU has been working on advancing the emotional intelligence of the computer – that is, teaching the machine to feel

NEW YORK, Oct, 26, 2017 /PRNewswire/ -- Last year's US presidential election results revealed the weakness of current projection methods. Oh Sang-gyoon, CEO of BPU Holdings, uses emotion-based intelligence to measure sentiment across the Internet to increase the accuracy of the predictive analytics. BPU worked with President Moon Jae-In during his presidential campaign to provide the team with issues-driven, real time dialogue powered by their Artificial Emotional Intelligence (AEI) engine, machine learning, and accurate sentiment analysis, allowing Moon Jae-In to respond and adjust to voter sentiment. Oh's analytics yielded estimates of voter turnout and projected election results that were accurate to within one percent.

The company has recently raised $60M, one of the largest crowd-funding success stories globally to date.

Korea IT Times spoke with Oh Sang-gyoon to talk about Artificial Intelligence and how BPU's pioneering AEI engine allows individuals and organizations to leverage social sentiment analytics at a whole new level, to potentially revolutionize other fields such as mental healthcare, education and law enforcement.

Q: What is AEI?

A: Emotional Intelligence (EI) is the ability of individuals to recognize their own and other people's emotions, discern between different feelings and label them appropriately, use emotional information to guide thinking and behavior, and manage and/or adjust emotions to adapt to environments or achieve one's goal(s). AEI is Artificial Emotional Intelligence – a similar kind of emotional intelligence exhibited by machines.

Q: How long has BPU been working to teach emotions to machines?

A: For the better part of the last decade we have been working on advancing the emotional intelligence of the computer – that is, teaching the machine to feel.

Q: Why an AEI OS?

A: As we carefully considered the approach desired to address the lofty objective of "teaching the machine to feel" we decided to build an open environment that allows a broad spectrum of applications and services access to AEI technology. This approach is perhaps more work than a closed single-purpose solution, but we believe it to be more useful and will have much broader and deeper impact on AEI development.

Consider the rate of technology adoption by humans. We have a hard time keeping track of our phones, let alone hundreds or maybe thousands of sensors and software agents that we are using and need to control. There is no doubt that in order to manage so much, we would need a computer's help. However, AI needs to be able to "adapt" and apply management according to human moods and behaviors without constantly bothering the human. If a computer has Emotional Intelligence, it can manage the world and respond to human events. With AEI OS, we take the concept of AI and apply it to each person. Everyone has an individualized AEI, many AIs all working together as a single AEI. So now you have multiple AIs (the AEI) all working on behalf of a single person, and millions of AEIs for everyone on earth. AEI OS is there to manage this complex environment securely and individually.

Q: Why would organizations care about emotional sentiment?

A: It is already a well-established fact that social and emotional learning (SEL) have a huge impact on education systems and businesses. Technology capable of teaching machines to exhibit these same qualities would be a powerful and useful tool benefitting many aspects of individual lives and the broader economy.

Beyond just the technical benefits, emotional wellness is key to human wellness. Organizations would benefit from implementing wellness programs that can provide measurable results for things such as healthcare, fitness, and human resources. Imagine being able to have a wellness app that can detect when an employee hasn't slept and is at dangerous fatigue levels; or someone who's been showing signs of stress and possible depression symptoms for over three months indicating a possible chronic problem. On a more positive note, imagine being able to offer awards and incentives for employees who actively pursue a healthy emotional wellness routine. Correlation can be shown linking these healthier employees to increased productivity levels, fewer sick days, and lowered cost of health care and down time. Organizations should be running, not walking, to adopt AEI programs.

Q: How has changing technology changed the way we view emotion?

A: Innovations like the printing press, the pre-paid postage stamp, the telegraph, the cell phone and the Internet are significant technologies that changed the equation of productive dialogue. Automation in communication innovation has evolved to facilitate and ease disruptive transitions in how we communicate and process information.

Enter the highly disruptive Computing Quadrangle. Mobile devices, social networks, global connectivity and invasive hacking by individuals and even governments make understanding sentiment difficult at best. Constituents simply do not behave as they did in the past. To measure accurate sentiment and to facilitate productive dialogue, a different approach must be used.

Q: Can you give me an example of that approach?

A: As it relates to emotional intelligence, words represent highly complex machine operations: open, close, move, etc. This is the job of the Operating System, to make the complex simple for the user. To this point, only rudimentary steps have been taken to assign emotion in these machine operations or to words in general.

AEI holds the promise of assessing emotional intent from words, paragraphs, and even visual cues like a furrowed brow or a smile. Hence, at BPU we are empowering the machine to sense emotion from written language the way we interpret body language. This is significant not just for a single application, but rather any application where reading emotional intent may influence our analysis. This requires an operating system on which to place a multitude of emotionally enhanced applications, making the complexity seem simple to the user.

The individual AEI also opens up an incredible new world of technology that must be personalized and customized to make it useful. Our heart rate, sleep patterns, and moods can all be adapted to and measured by emotionally intelligent technology. Modern businesses that implement AEI can now collaborate to maintain a complete profile of the person and what it takes to make his or her life easier. AEI will be used everywhere, in our cars, homes, offices and on our smart devices. Imagine the day that when you open a new phone and connect to the Internet, all of your personalized data is imported by AEI.

Q: What kind of organizations would be able to use this technology?

A: What organization would NOT benefit from enhanced interpretation and analysis of the mountains of available information online? The "big data" age is upon us and we must develop new approaches to maximize the benefits. Small organizations benefit from being able to increase efficiency that will help the business grow. Large organizations, because of their complexity, benefit because AEI OS will not forget about the individual and will help the organization understand human patterns within.

Q: Why are regional and geographic language differences difficult for the machine to interpret?

A: Put simply, you can't take the one language and map it directly to another language. There are subtle nuances within different languages that differentiate levels of endearment or intensity. While emotions are pretty standard across human history, the language used to express them is anything but standard.

Q: Does your engine understand emojis? 

A: Yes. Not only does the AEI OS accurately measure emotion from emojis, the system is learning to understand visual cues in photographs and other graphics. It is exciting the watch the system teach itself.

SOURCE Korea IT Times