Written by Lin Lingyi, KrAsia
When Dennis Ye was five years old, he asked one question that made a difference for his grandaunt, who had been diagnosed with cancer and had not laughed in an entire year.
“One day, we visited them at their house. When I saw a huge diamond ring on her finger, I childishly asked, ‘Is it real?’ She immediately laughed,” Ye told KrASIA, adding that his granduncle then increased his red packet from SGD 50 to SGD 500, because of the momentary joy he brought to the household.
This experience has in some way influenced his career, Ye said. In 2019, along with entrepreneurs Bai Jiamin, Teodor Andius, and Nini Chang, he founded deep-tech startup Reality Detector, a Singapore-based firm that produces video-based deception software that identifies deception cues in videos.
The software can detect facial and body signs suggesting that an individual may be lying, from video recordings and without requiring physical contact, by utilizing AI, deep learning, and computer vision methods. The startup says that its software solution can be used to validate interviewees’ responses, or even ensure that individuals declare their travel history truthfully, “especially in the world of the post-COVID new normal.”
“When I worked in an accelerator, and later as a financial regulator, we interfaced with many startups. I was concerned about people generally, who may defraud others’ time and money,” explained Ye.
In October 2020, Reality Detector scored SGD 500,000 (USD 368,000) in funding from Silicon Valley’s Draper Associates, KrASIA reported.
Ye, who earned his master of business administration at China’s Tsinghua University, highlighted the influence of guest professor Peter Thiel, co-founder of PayPal and Palantir Technologies, and one of the first investors in Facebook, for inspiring him to turn towards technology to start his venture.
|Teodor Andius (left) and Dennis Ye (right), two founding members of Reality Detector|
An alternative polygraph
Historically, polygraphs, which are devices used for lie detector tests, were pilloried for being ineffective, inefficient, and overly intrusive, as they used three basic indicators of autonomic arousal: heart rate or blood pressure, respiration, and skin conductivity. However, Reality Detector emphasizes that its software represents a better alternative rooted in behavioral science and deep learning to identify deception cues.
Reality Detector obtained grants during its early-stage from the Agency for Science, Technology, and Research (A*STAR) as well as Enterprise Singapore, both statutory boards under Singapore’s Ministry of Trade and Industry. “The support from Enterprise Singapore and A*STAR were indispensable for covering costs and building technological innovation respectively,” he said.
With the video surveillance market growing at a compound annual growth rate (CAGR) of 14.2%, the firm wants to prove its technology in industries where sensors, cameras, and data analytics software are in high demand.
“There are multiple carefully selected scenarios where the pros outweigh the cons of employing these tools. For example, in testing whether drug mules are lying or framed, or preventing the spread of COVID-19 by identifying people who lie about travel histories,” Ye said to KrASIA while demonstrating the current version of his product.
Ye suggested that the software should first be used in credibility assessments, rather than being used as standalone legal evidence. Currently, Reality Detector’s first product is in late-stage alpha testing and is close to securing its first customers.
“Tim Draper [founder of Draper Associates] gave us a piece of advice. It probably won’t work too well at the start, but then it will probably start working too well. We need to think about whether the product is ethical and legally robust and be cautious about possible claims we make about our technology. Our products need to go through many rounds of validation against real-world data.”
The technology behind alternative lie detectors, however, is still immature. Other companies, such as EyeDetect, have come under scrutiny for unreliability and inaccuracy. In 2018, the polygraph industry was worth USD 2 billion in the US, where about 2.5 million polygraph tests are conducted every year.
|Sample interface of Reality Detector’s product. The actual interface is in the process of being patented.
Courtesy of Reality Detector
Building Singapore’s entrepreneurial system
Despite recognizing the pain points of his industry, Ye has grand plans for his firm. “I intend to lead the company to an IPO,” he said.
“I was advised by a friend in China that starting a competitive AI company in Singapore was foolish because China has companies like Megvii and SenseTime. But we have a chance as this field of psychology is not well known in China. Still, it will be difficult,” Ye said.
The firm has a respectable team of advisors and investors, including Tim Draper, who invested in SpaceX and Baidu; Andy Lim, chairman of Tembusu Partners; and Shirley Wong, managing partner at TNF Ventures. Ye himself comes from a line of entrepreneurs—his great-grandfather, Tan Cheng Siong, contributed to the establishment of Oversea-Chinese Banking Corporation (OCBC) in Singapore.
“There is so much we want to do for the community. We want to help create good jobs, and ensure Singapore is at the top of the value chain for technology,” he affirmed.
This article was written by Lin Lingyi and is part of KrAsia’s “Startup Stories” series, where KrAsia writers speak with founders of tech companies in South Asia and Southeast Asia. It is republished here with the permission of KrAsia. For more Asia business and technology news visit: https://kr-asia.com