Information & Engineering Breakthroughs: Network Coding, the Shannon Award, and from MMLab to SenseTime

# Information & Engineering Breakthroughs: Network Coding, the Shannon Award, and from MMLab to SenseTime

CUHK engineering research boasts two world-class high-water marks, each with a radically different character: one is Network Coding, pioneered by Raymond W. Yeung — a new field of information theory in which a CUHK scholar served as co-founder, culminating in the 2022 Claude E. Shannon Award, the discipline's highest honour; the other is Tang Xiao'ou's Multimedia Laboratory (MMLab) — which made landmark contributions at the inflection point of the deep learning wave in 2014 and directly spawned the AI unicorn SenseTime. The former "rewrote the knowledge of a field"; the latter "incubated laboratory results into industry." This article is a factual archive for the Reference Zone (04 Research). No credibility badges are assigned. Each item is supported by academic/official/secondary sources. The scholars discussed are public figures, identified by name according to public records. (Tang Xiao'ou passed away in 2023.)

# Part I · Network Coding: A New Field of Information Theory Pioneered at CUHK

# 1. What is Network Coding: A Counter-Intuitive Insight

In the traditional view of network communication, a node in the network does only one thing: store and forward — receive a data packet, retransmit it intact, like a post office sorting letters. Network coding proposes a counter-intuitive insight: intermediate nodes can "encode" (combine via computation) multiple received data packets before sending them onwards, and the receiving end can then "decode" them to recover the originals — which can, in certain network topologies, significantly improve transmission efficiency.

According to the IEEE Information Theory Society's Shannon Award citation^※ and the English Wikipedia^※, network coding theory was co-founded by Raymond W. Yeung and collaborators in the late 1990s, and it "fundamentally changed the understanding of network communication." The foundational paper "Network Information Flow"^※ was published in 2000 (co-authored by Yeung with R. Ahlswede, Ning Cai, and Li Shuo-Yen). The classic demonstration of this theory is the "butterfly network" — a topology where a bottleneck would occur if intermediate nodes only forward data; if encoding is allowed, it can simultaneously satisfy the demands of multiple receivers, approaching the network's capacity limit.

A paradigm shift: The significance of network coding lies not in inventing a specific protocol, but in redefining "what a node in a network can do" — elevating it from a passive "courier" to an active "computational agent." This represents a major expansion of information theory in the "network" dimension since the days of Shannon.

# 2. The Founder and CUHK's Role

According to the English Wikipedia^※ and a CUHK official press release^※, Yeung is the Choh-Ming Li Professor of Information Engineering at CUHK and Co-Director of the Institute of Network Coding at CUHK. He is a co-founder of the field of network coding. This means: network coding is not research that a CUHK scholar merely "participated in," but a completely new field where CUHK served as one of the core birthplaces. The Institute was established in 2010, allowing the theory to be continuously developed for practical scenarios such as data storage and network communication.

# 3. Theoretical Depth: From the Zhang-Yeung Inequality to BATS Codes

Beyond network coding, the team has made multiple contributions to the fundamental theory and applications of information theory. According to the English Wikipedia^※:

• The Zhang-Yeung Inequality: Yeung, together with Professor Zhang Zhen of the University of Southern California, discovered this inequality, establishing the existence of a completely new class of "non-Shannon-type information inequalities." This was a major breakthrough in foundational information theory, as it was previously believed that Shannon-type inequalities had exhausted all constraints governing information measures.
• BATS Codes (BATched Sparse codes): He invented BATS codes, which can improve transmission rates over networks subject to packet loss, a representative achievement in translating network coding theory into engineering practice.
• Global Impact of a Textbook: The information theory textbook he authored is reportedly used by over 100 universities worldwide — meaning he has not only pioneered a field but also shaped the training of a generation of information theorists.

The "Foundation × Application" Twin Engines: The Zhang-Yeung inequality belongs to the most abstract realms of basic theory (mathematical constraints on information measures), while BATS codes directly tackle a highly practical engineering problem (transmission over lossy networks). The ability of a single team to exert force simultaneously at the theoretical ceiling and the applied floor is characteristic of top-tier information theory research.

# 4. Recognition: The 2022 Claude E. Shannon Award

According to the IEEE Information Theory Society^※ and a CUHK official press release^※, Yeung received the 2022 Claude E. Shannon Award, recognising his "consistent and profound contributions to the field of information theory," and delivered the Shannon Lecture at ISIT in Finland that year. Named after the founder of information theory, the Shannon Award is the highest honour in the field, approximately equivalent to the discipline's "highest lifetime achievement award." Furthermore, according to CUHK and IEEE records (see also UGC Awards^※), Yeung previously received the 2016 IEEE Eric E. Sumner Award and the 2021 IEEE Richard W. Hamming Medal. A CUHK scholar receiving this honour marks the University's world-class standing in the foundational discipline of information theory.

# Part II · From MMLab to SenseTime: A Legend of Deep Learning Industrialisation

# 5. Asia's Only AI Lab Ranked Among the Global Top Ten

The core vehicle for CUHK's world-class standing in artificial intelligence (particularly computer vision) is the Faculty of Engineering's Multimedia Laboratory (MMLab). According to the English Wikipedia entry for "Tang Xiao'ou"^※, MMLab was founded by Professor Tang Xiao'ou of the Department of Information Engineering and was, at the time, the only research lab in Asia among the world's top ten artificial intelligence laboratories. It was also the sole Asian representative on Nvidia's 2016 list of the world's ten leading AI labs.

In a landscape where AI research was heavily concentrated in North America, a Hong Kong university lab's inclusion among the "global top ten, Asia's only" is itself powerful testimony to CUHK's engineering research strength. This laboratory was the wellspring of the subsequent SenseTime legend.

# 6. DeepID: The Face Recognition System That Reportedly First Surpassed Humans

MMLab's most iconic achievement is the DeepID face recognition algorithm from 2014. According to English Wikipedia, "SenseTime"^※ and the DeepID academic paper (CUHK EE)^※, in 2014, the team released the face recognition algorithm DeepID, acclaimed as the world's first algorithm to exceed human-level accuracy in face recognition. This breakthrough, based on deep learning technology, marks one of the landmark moments in the history of computer vision development.

The significance of "surpassing humans": Face recognition was long considered a task where "machines could hardly match humans." DeepID reportedly crossed this threshold in accuracy for the first time, symbolising a qualitative leap in computer vision capability during the deep learning era. The CUHK team stood at the very forefront of this global technological inflection point.

# 7. SenseTime: From Laboratory to Industry

The technological breakthrough of DeepID quickly translated into industrial force. According to English Wikipedia, "SenseTime"^※, in October 2014, SenseTime was co-founded by Tang Xiao'ou and computer scientist Xu Li, among others. It is a research-driven spin-off incubated from the academic research at CUHK's MMLab and subsequently grew into one of the most representative companies in China's computer vision sector. According to the source, the CUHK-SenseTime Joint Lab has published or presented over 400 computer vision papers in top global academic journals and conferences — a volume second only to Microsoft. This means the "CUHK-SenseTime" academic-industry combination ranks highly in global academic output within computer vision.

The "Academia-Industry Integration" paradigm: The SenseTime story is a complete closed loop of "laboratory breakthrough → spin-off enterprise → joint lab feeding back into academia" — the lab produces technology and talent, the company provides application scenarios and resources, and the joint lab channels applications back into new academic results.

# 8. A Necessary Boundary Clarification

As a listed company, SenseTime's operations, valuations, and commercial disputes are corporate matters subject to flux with market and time; this archive does not track or judge them. This article only records two traceable facts: the academic achievements of CUHK's MMLab and the incubation arc of SenseTime as an academic spin-off. Conflating a company's commercial performance with a laboratory's academic contributions is neither accurate nor this archive's remit. Readers concerned with SenseTime's commercial situation should consult its financial reports, regulatory disclosures, and other primary information.

# Part III · The Convergence of Two High-Water Marks

When viewed alongside network coding and DeepID/SenseTime within the broader genealogy of CUHK research, they join optical fibre and NIPT to form a landscape of "original breakthroughs + far-reaching impact," each with distinct emphases:

If optical fibre, network coding, and NIPT are more about "rewriting the knowledge of a field," then DeepID/SenseTime more vividly demonstrates the path of "directly incubating laboratory results into industry," echoing the collaborative logic of "Hong Kong basic research + mainland China industrial translation" within the Greater Bay Area. The deepest industrial impact of a university sometimes lies not in how many companies it starts, but in whether one of its laboratories happened to stand at the inflection point of a technological era — and during the 2014 deep learning wave, CUHK's Multimedia Laboratory was right there, making significant contributions.

Related reading: Overview of Research Achievements, Research Output and Spin-offs, State Key Laboratories, Life Sciences and Medical Breakthroughs.