Taiwanese start-ups pushing ahead with AI tech 台灣十家最酷科技新創 AI安全帽攻進歐美

Tue, Apr 10, 2018 - Page 14

The Ministry of Science and Technology last week published a list of the “10 coolest Taiwanese tech start-ups” for the year 2017. The list was made up of five companies involved in artificial intelligence (AI), two in the Internet of Things, a further two in biomedical sciences and another working with augmented and virtual reality.

One of the companies on the list is JARVISH, founded by Lu Tsung-hsien, which in February last year announced the world’s first mass-produced “smart” safety helmet. Lu says the company carried out R&D, built a team and made its product all within Taiwan. According to Lu, the main advantage of their product is it has AI technology embedded within the helmet which is able to detect if a motorcycle has crashed and then, via a smartphone app synchronized with the helmet, use conversational robotics to initiate a handsfree emergency call.

Meanwhile, in a television report recently broadcast by the BBC called “In Your Face: China’s all-seeing state,” a reporter was allowed access to a surveillance center in China’s Guizhou Province to carry out a real-time experiment. After the journalist’s face was scanned and registered on its system as a suspect, it took only seven minutes, walking from Guiyang city center to the city’s train terminus, for the journalist to be apprehended by Chinese police. The report has sparked intense debate online.

In Taiwan, every residential area contains security surveillance systems, and Taiwanese start-up Umbo CV has developed its own image recognition system. Director of design Chen Chao-yi says the company’s Umbo Light product, which uses deep learning and neural network architecture to commercialize computer vision on security cameras, was developed using Umbo CV’s in-house AI image recognition system. The product uses human-based knowledge assessment systems to determine whether surveillance images contain abnormal objects or behavior, and will issue an alert immediately an incident is positively identified.

Chen says that in the past, surveillance systems would only judge the movement of objects, but Umbo Light also uses objects’ behavioral patterns to determine whether a security problem exists. After performing an analysis, if behavior such as wall climbing, breaking in or the shadowing of an individual is detected, an alert is instantly sent to the client.

Chen explains the technology is like having an experienced AI security guard on duty 365 days a year, 7 days a week and will dramatically reduce the human resource cost of security, which was previously reliant on the human eye.

(CNA, translated by Edward Jones)

科技部上週公布「二○一七年台灣十家最酷科技新創」名單,其中屬於人工智慧的公司計有五家,物聯網及生醫各兩家、AR/VR 一家。

其中,酷設工坊創辦人呂宗賢說,他們是台灣研發、團隊、製造,去年二月底發表全球第一個成功量產的智慧安全帽,最大的優勢是利用人工智慧放入安全帽中來偵測機車摔倒事件,透過智慧型手機APP與安全帽的同步連線,利用對話機器人來達到免手持的緊急電話。

之前英國廣播公司播出,「中國全程監控你的臉」的電視報導讓記者在貴州監控中心實地測試,結果他的臉部被掃描登錄為嫌疑犯後,從貴陽市區走到車站僅七分鐘,就被大陸警察攔獲,引起網路上的討論不斷。

台灣的居住環境都有安全監視系統。研發影像辨識的盾心科技設計總監陳昭億說,Umbo Light是由他們自主研發的人工智慧影像辨識系統,運用深度學習的類神經網路架構,將電腦視覺商用化在安全監控相機上,透過人類的知識教導監視器主動判斷畫面中的物體與異常行為,並在事件發生第一時間進行通報。

他解釋,過去監視系統只會對物體的移動進行判斷,Umbo Light則是利用生物體的行為模式去判斷有無安全疑慮,經過分析,若出現爬牆、入侵、尾隨的行徑,將第一時間通報給客戶傳達警訊。

陳昭億解釋,就像是一個全年無休的優秀AI警衛,可以大幅減省以往用人眼盯梢,所需耗費的鉅額人事成本。

(中央社)

讀後練習

follow up

During a recent interview, billionaire tech entrepreneur Elon Musk said he believes artificial intelligence is more dangerous than nuclear weapons:

“I am really quite close, I am very close, to the cutting edge in AI and it scares the hell out of me. It’s capable of vastly more than almost anyone knows and the rate of improvement is exponential.

“So the rate of improvement is really dramatic. We have to figure out some way to ensure that the advent of digital super intelligence is one which is symbiotic with humanity. I think that is the single biggest existential crisis that we face and the most pressing one.

“And mark my words, AI is far more dangerous than nukes... So why do we have no regulatory oversight? This is insane.”

Questions

1. Do you agree with Musk that AI has the potential to evolve into a threat more dangerous to humanity than nuclear weapons?

2. Musk has called for regulatory oversight of artificial intelligence. Do you think this is a good idea — or even possible?

3. Are you worried by the number of surveillance cameras in operation today? Are we sleepwalking into becoming a “surveillance state”?

(Edward Jones, Taipei Times)