The 2019 World Innovators Meet (WIM 2019) kicked off on December 06 in the heart of Beijing. The three-day-long event, hosted by EqualOcean, brought together more than 200 speakers from twenty different countries, and over 6,000 innovators, to discuss recent innovation trends in world tech.

During the conference, nine forums were organized around the hottest topics in relevant industries, including new retail consumption, AI, fintech, automotive & mobility, semiconductors, e-commerce, real estate, smart manufacturing and other sectors. 

EqualOcean released a series of research reports on the new generation of companies across several sectors around the globe, including "Artificial Intelligence: Trends and Startups 2019," "EdTech: Trends and Startups 2019," "Automotive and Mobility: Trends and Startups 2019," Enterprise services: Trends and Startups 2019" and others. You can find these reports both in the English or Chinese versions here.

At the Innovation Leaders’ Summit held on December 8, Jun Zhou (周军), the founder and CEO of Eyecool Technology (眼神科技), delivered a keynote speech, with "Multimodality" as the keyword. 

Company overview

Beijing-based AI-driven company, Eyecool Technology (眼神科技), focuses on deep learning technology, relying on big data and biometrics as the entry point. It has become among the few companies in China that offer different types of biometric technology services, including fingerprint, face recognition, vein recognition, and iris recognition. 

The company claims that in its 20 years of existence it has served more than 2,000 customers worldwide such as key domestic banks: China Construction Bank (中国建设银行), Bank of China (中国银行), Agricultural Bank of China (中国农业银行), Bank of Communication(交通银行), and some key domestic universities such as Tsinghua Univesity (清华大学) and Shandong University (山东大学). 

Founded in 2016, Eyecool has raised a total of over CNY 100 million, mainly led by Beijing-based investment firm UCF investment (先锋国盛基金).

What is "Multimodality"?

The conception of multimodality is a new term from the field of communication and expression. It describes the communication practices in terms of textual, oral, linguistic, spacial, and visual resources or modes used to compose messages.

According to Zhou, Eyecool has spent more than 20 years on research and development in the field of biometrics, all with the aim of enhancing multimodality, using artificial intelligence as the backbone and finally realizing an "AI-driven Multimodality biometrics" model for adoption and innovation. 

Multimodality is not just a mere additive combination or a simple combination of semiotic modes, but it is based on the merging of data and algorithms, supported by the decision platform, Mr. Zhou emphasized.

How to solve the problem of misleading identification and sensor perception limitations, and how to make machines perceive and understand the world like humans are areas that could be facilitated by multimodality, according to Mr. Zhou.

Multimodality draws the attention of tech giants

The fast-growing demand in the global biometrics market has driven the attention of well-known Chinese tech giants such as Huawei (华为), Baidu (BIDU: NASDAQ), Tencent (TCEHY: OTC) and others. They have become leading investors in the biometrics market in China, said Mr. Zhou in his speech.

In early June this year, according to the UK-based news portal Planet Biometrics, Huawei acquired the Russian-based facial recognition tech firm Vocord for USD 50 million. Despite being blacklisted by the US authorities, the tech giant has defied all the odds and is among the top smartphone producers in 2019. 

Recently, Baidu formed a partnership with Intel to collaborate on hardware and software platforms for artificial intelligence to improve biometric facial and voice recognition performance, while Tencent developed an image-aging technology for facial recognition of missing children, which is planned to be used by police searching for missing or abducted children.

From weak artificial intelligence to general artificial intelligence

Despite the recent achievements in the field of AI, and increasing awareness of how it can be used in all industries, artificial intelligence is still in a weak stage of development. Dr. Zhou explained this aspect of how "Multimodality" can break the  barrier between weak and general AI; hence, multimodality is an essential step in the move towards general artificial intelligence.

For many observers, general artificial intelligence is a direction that needs to be taken in an era of weak artificial intelligence. Weak artificial intelligence covers "a wide range of knowledge" and solves the problem of perception, while general artificial intelligence has several superior attributes: self-learning, incremental learning and enhanced learning. These contents are combined into a strong artificial intelligence. The singularity that separates weak artificial intelligence and strong artificial intelligence has not yet been reached. 

The era of strong artificial intelligence that "sees less and know more" and "sees before you see" is something that many cutting-edge international artificial intelligence companies have been working on for many years, and technology has hit the ceiling.

Foreigners companies adhere to the multi-modal, as the market grows strong

In recent years, the biometrics market has been experiencing rapid growth worldwide, with major foreign companies such as NEURO Technology, CROSSMATCH, IDEMIA MORPHO and DERMALOG all striving to increase the status of multimodality.

By 2024, the global biometric market size is expected to reach USD 65.3 billion, from USD 33 billion in 2019, with a CAGR of 14.6%, according to data released by MarketsandMarkets.

How the tech companies will achieve a breakthrough and realize a general artificial intelligence is the ultimate classroom "quiz" that the R&D departments are trying to solve.

However, "multimodality" might be the key to solve such a problem in the long run. More innovators are joining the field and taking their own directions in the study of multimodality instead of the single modality. This will create a suitable environment for the future development of artificial intelligence.