(Original Title: What Can Be Done to Solve the Problem? Voice Interaction and Face Recognition Expected to Pave the Way for AI Commercialization)
Editor's Note
[Artificial intelligence (AI) has advanced significantly today, especially over the past few years. It has made remarkable strides and is increasingly integrated into our daily lives, particularly in voice interaction and facial recognition. These technologies, once seen as futuristic concepts, are now part of our reality. Imagine what the AI landscape might look like in a decade—just as smartphones revolutionized our lives a decade ago, how will AI transform the way we live?]
While many may not yet fully grasp the transformative trends, others have already begun to capitalize on them, such as venture capitalists. Capital is the most sensitive indicator of AI’s potential, and investment in AI companies has been rising annually. Last week alone, several domestic AI-related companies secured funding. It’s anticipated that 2017 will mark the beginning of global AI commercialization.
â— Competing for Intelligent Voice Interaction Consumer Services
On July 5th, Alibaba officially launched its first smart speaker, the Tmall Genie X1. This followed Amazon Echo’s impressive sales of nearly 10 million units. Smart speakers have become a crucial carrier for voice interaction technology, thanks to their potential as the control hub for various household devices. Furthermore, voice assistants like Apple’s Siri, Microsoft’s Xiaobing, and Duowei are paving the way for real-time voice translation, voice input, and voice-controlled systems, opening up numerous application fields and accelerating consumer-grade services.
In automotive applications, Hong Kong University of Science and Technology has collaborated with Mercedes-Benz, BMW, Toyota, and other international manufacturers, as well as domestic carmakers. Earlier this year, Hong Kong University of Science and Technology and Changan Automobile announced comprehensive cooperation in intelligent technology development, product design, and vehicle applications.
Baidu’s founder, Li Yanhong, emphasized in speeches that the integration of hardware and software in AI products will be a key direction. In February, Baidu acquired Raven Technology. Externally, Baidu is stepping up the application of AI technology through the integration of hardware and software, ensuring that products are integrated into everyday life.
Since its debut earlier this year, Baidu DuerOS has partnered with renowned companies like Midea, Haier, TCL, Lenovo, Vivo, HTC, CITIC Guoan Guangshi, and Xiaomi. It has also ventured into television, refrigerators, smart wearables, and automotive scenarios, offering users a seamless living experience.
In an interview with the reporter of the "Daily Economic News," Longbirth, Chief Marketing Officer of Stables, expressed optimism about vertical field applications, focusing on smart homes, vehicle IoT, robotics, and other areas to drive industrial upgrades through traditional technology companies.
â— "Invisible" Facial Recognition
Voice technology has gained widespread adoption through smartphone assistants and smart speakers. Image recognition, however, remains less discussed by consumers. Does this mean image recognition is distant from us?
In an interview, Long Mengzhu noted that users employ AI image technology in selfie apps like Meitu Xiuxiu, where capturing images requires automatic face detection. Beauty enhancements and special effects require precise face, nose, and eye locking. Baidu Maps and Tmall's product search are both powered by image recognition technology. Voice technology feels more intuitive because users actively engage in speaking and receive immediate feedback.
Dr. Xu Li, co-founder and CEO of Shang Tang Technology, told the "Daily Economic News" reporter that facial recognition has been widely used in public security surveillance systems, home security systems, identity verification, and entertainment internet platforms, making it the leading application in computer vision. Shangtang Technology’s clients include many banks and financial firms, collaborating with JD.com’s wallet service, allowing users to scan their faces to unlock passwords instead of traditional methods.
Jeff Lin, founder of computer vision startup Insight, noted in interviews that facial recognition is currently one of the hottest areas for AI startups, particularly in financial identity authentication and security scenarios.
Unlike voice interaction, computer vision companies typically provide technical support to other businesses rather than directly selling products to consumers. Dr. Xu Li explained that while current technology can exceed human accuracy in specific visual tasks, the overall video scene content and object recognition remain below human standards, explaining why consumer-level computer vision products are scarce.
◠China’s Opportunities
Human-computer interaction through "intelligent voice operating systems" and "facial recognition" as the core, along with interconnected smart homes, IoT, security, and identity recognition scenarios, are gradually becoming the primary battlegrounds for AI commercialization. To achieve vertical applications in areas like smart audio and facial recognition, AI startups are exploring more use cases.
Wang Shangzhi, a computer industry analyst, told the "Daily Economic News" reporter that China has many advantages in AI, expecting its commercialization to surpass global standards. The vast market potential and high population density determine the demand for personalized services. Additionally, AI is expected to replace repetitive low-end labor.
"In terms of data and business scenarios, China inherently has advantages. With a large population, technology companies can easily access frontline data across various scenes. Secondly, China offers all the application scenarios for AI field practices, and the environment is receptive to new technologies. Thirdly, China has sufficient reserves of AI engineers. Based on these three conditions, Chinese AI can lead globally," Xu Li told the "Japan Economic News" reporter.
However, Wang Shangzhi pointed out that China’s entrepreneurial environment and mechanisms are not yet mature, and laws and policy supervision are not yet perfect, hindering the rapid integration of AI technology with practical applications.
[No More Investments Before First Investment in Past Week for at Least Four AI Companies]
Every reporter Zhang Si edited by Xi Jinpi
Capital is the most sensitive indicator, and AI investment companies are increasing annually. Last week, Chinese startups including Tang Focus Technology, robotics company Geek+, cloud-brain technology, and Tesla Technology announced financing information. Shangtang Technology’s $4.1 billion Series B round set a record for the global single-wheel AI market.
Multiple investors told the "Daily Economic News" reporter that with the accumulation of big data and improvement of cloud computing infrastructure, AI technology is bringing tremendous changes to industry development, and 2017 is expected to become the first year of global AI commercialization.
"Artificial Intelligence +" Becomes a New Trend
What sparked the sudden surge in AI interest? Almost everyone interviewed by this reporter attributed it to "AlphaGo."
The wave of artificial intelligence has developed for five years so far. Lan Chi Ventures has been focusing on AI since 2012. They concluded then that the foundation of AI is big data, and the foundation of big data is cloud computing. Solving the issues of data and underlying cloud computing platforms is urgent.
Cao Yu, executive director of Lan Chi Ventures, told the "Daily Economic News" reporter that Baidu founder Li Yanhong's push for autonomous driving on the fifth ring is an industry signal. The delivery of autopilot technology through real-world operations indicates that around the accumulation of technology, multi-scenario applications are now feasible. AI development has reached the tipping point for commercial breakout.
In an interview with the "Daily Economic News," Lenovo's Star Investment Manager Wang Zheng discussed Shang Tang Technology's financing case. He believes the high financing stems from the company's extensive applications. Early on, Shang Tang Technology had a strong ability to identify industry trends. After accumulating operational data, it became clear about application scenarios.
Lenovo Star is one of China’s major AI investment institutions. Its early investments in Contempt Technology and Incischi are expected to become unicorn potential companies in computer vision and speech recognition.
These two investment institutions represent just a microcosm of the massive investment network. This March, AI was written into the government work report. National leadership attaches great importance to AI as a strategic development technology. Domestic IT giants like Baidu, Alibaba, Tencent, and international tech giants like Amazon, Apple, Facebook, and Microsoft all see AI as the breakthrough point for the next technological revolution, investing heavily to accelerate deployment.
According to CB Insights statistics, 61 AI enterprises received funding in 2010. By 2016, 522 AI startups had secured financing, with investment soaring from $81 million to $3.12 billion.
Why does the market place such emphasis on AI development? According to Accenture's research report, comparing the economic growth rates of 16 major industries, industries integrating AI technologies show higher growth potential. It is estimated that by 2035, AI can increase the average profitability of involved industries by over 30%, boosting the total economic value by nearly $14 trillion.
There is no doubt that "AI+" will become a new trend, not a specific track for O2O, e-commerce, or consumer upgrades, but a technological tool for industrial upgrading—like the Internet, enabling O2O, e-commerce, and consumer upgrades. When AI matures, all industries can use it to enhance efficiency and productivity. This explains why it garners so much attention from capital and the industry.
How High Financing Leads to High Valuation
Academically speaking, investigating valuations is the basic logic for some AI sector investments in the past, but this situation is changing this year.
"Investment is a risky business. Since it is a business, it must pay serious attention to investment. Especially for technical projects, it is more difficult to realize. Only companies emphasizing technology and not landing will be treated cautiously," said Wang Zheng. Companies achieving high valuations focus not only on team backgrounds but also on their ability to obtain various resources, including selecting the right track.
Cao Yong shares the same view, "Under the accumulation of basic technologies, we must choose a large industrial track, such as logistics. Besides team background and technical capabilities, we must understand their comprehension of the logistics industry. Whether they can truly help logistics companies achieve spatial optimization through data processing requires scientists and tech founders to understand industry needs and ensure their designs meet both industry applications and sales capabilities to ensure enterprise usage."
According to Zhiyan Consulting's research report, areas receiving large investments include natural language processing, robotics, smart homes, smart driving, and smart finance. Among these, natural language processing, robotics, and computer vision have received substantial investments, all exceeding 1.5 billion yuan.
In addition, according to Accenture's research report, anthropomorphic technologies are expected to be widely used in communications, manufacturing, and finance by 2035, becoming the top three industries with the highest economic growth.
However, the large-scale application of AI still lacks successful cases. AI enterprises struggle to realize commercialization. Behind high investments, AI enterprises are still exploring commercialization scenarios. Without combining technology with practice, AI risks becoming a bubble waiting to burst.
Moreover, premature business concerns also hinder academic progress in AI. The industry believes the application bonus brought by the last breakthrough in academia (deep learning algorithm) may soon be exhausted.
[AI Talent Demand Increases Eightfold in Three Years]
Each journalist Zhang Si intern reporter Zong Xu edited by Xie Jinchi
According to the "Global AI Field Talent Report" released by LinkedIn, demand for AI talent has surged in the past three years. Since 2014, AI-related job postings on the LinkedIn platform have grown from approaching 50,000 to over 440,000 by 2016. The rapid increase in job demand not only reflects AI's swift development but also highlights the shortage of talent.
Why is the demand for AI talent so high? Industry insiders interviewed said that acquiring talent means securing the future. In the next 5 to 10 years, the shortage of AI talent may worsen.
Talent flows into normalcy
The scarcity and high mobility of personnel have driven the rapid rise in AI talent prices. In this "arms race" for AI talent, offering high salaries to attract senior experts and university researchers has become a common tactic among high-tech companies.
In 2013, Google invited Geoffrey Hinton to join the Google Brain team. Hinton’s company DNNresearch had no product and only three employees. Similarly, Apple acquired Turi, an artificial intelligence company founded by Guestrin in Seattle, in 2016.
Additionally, the surge in talent demand has led technology companies to focus on top academic talents in major universities. In 2013, Facebook hired Yann LeCun, a professor at New York University and one of the three leaders in deep learning, as the head of the Facebook Artificial Intelligence Lab; Google invested $4.5 million at the University of Montreal; and Intel donated $1.5 million to establish a machine learning and cybersecurity research center at Georgia Tech University.
In China, the most famous AI talent acquisition case was Wu Enda joining Baidu. Wu Enda joined Google X Lab in 2010 and was responsible for creating the world's largest neural network, "Google Brain," in 2011.
However, under the AI storm, the acquired talents may not stay. For instance, after Baidu successfully recruited Wu Enda, Executive Vice President Yu Kai announced his departure in June 2015. In March this year, Wu Enda also announced his resignation from Baidu. Senior Vice President Wang Jin, formerly the general manager of the autonomous driving department, also announced his departure. Additionally, Zhang Qian, who had previously left, and other top AI talents have departed. In the domestic AI talent market, Baidu plays the tragic role of the "Huangpu Military Academy."
Why do these companies spend heavily on talent? Dr. Xu Li, co-founder and CEO of Shang Tang Technology, told the "Daily Economic News" reporter that acquiring talent is about securing the future. "When Google paid £400 million for the DeepMind company, only 50 people worldwide truly understood deep learning engine designs, with 12 of them at DeepMind. So Google essentially bought future engine designs."
Lack of Creative Talent
For AI talents, Dr. Greg S Corrado, a senior researcher in the field of artificial intelligence and machine learning at Google, believes there are two types: engineers who genuinely understand and use AI tools and innovative, business-savvy individuals who can effectively apply AI technology more broadly.
According to the Global AI Field Talent Report, the number of global AI talents exceeded 1.9 million in the first quarter of 2017, with the U.S. having over 850,000, occupying half of the global total; China had 50,000, ranking seventh in the world. In terms of talent structure, there is a significant gap between the number of senior AI talents in China and the U.S.
Computer industry analyst Wang Shangzhi analyzed in an interview with the "Daily Economic News" reporter that in the field of AI, China has a strong talent pool, particularly in the quality of "engineers," but China still lacks "creative talents." "There is a lack of profound understanding of business model changes and global influence."
Xu Li also believes that China indeed has a large reserve of future AI engineering personnel but lacks top-tier research talents. "Above all, AI research can truly guide the promotion of talents in the entire industry. It is very scarce. The two countries are relying on the smartest people to compete."
Professor Deng Zhidong of Tsinghua University analyzed in an interview with the "Daily Economic News" reporter that with the rise of the industry, the shortage of AI talents may increase in the next 5 to 10 years.
How to Address Talent Cultivation? This issue involves three parts: government, business, and university. According to Deng Zhidong, as the government, it is necessary to introduce policies, such as supporting the establishment of more AI majors in first-class universities, creating AI colleges, or supporting social education through online teaching methods like MOOCs, and innovating institutional mechanisms to cultivate a large number of AI R&D talents for society. "The University of Chinese Academy of Sciences has recently established an AI Technology College, which is a good start."
Chinese universities and research institutes should pay more attention to exploring and innovating cutting-edge technologies for AI, cultivating a large number of engineering and technology development talents for Chinese AI companies. Moreover, enterprises must accelerate the process of industrial landing and realize their commercial value.
Stage Led City Lights,City Led Lights,Man City Led Light,Smart Urban Lighting
Guangzhou Cheng Wen Photoelectric Technology Co., Ltd. , https://www.cwledwall.com