Features

The PLA’s AI Competitions

Recent Features

Features | Security | East Asia

The PLA’s AI Competitions

Can the new design contests foster a culture of military innovation in China?

The PLA’s AI Competitions

A child watches a video depicting the flow of digital information during the National Science and Technology Week exhibition held at the Military Museum in Beijing on May 24, 2019.

Credit: AP Photos/Ng Han Guan

In recent years, the People’s Liberation Army (PLA) has sought to facilitate innovation through hosting “Artificial Intelligence (AI) Challenge”-style competitions. Interesting content aside, these PLA equipment departments-led competitions essentially serve as a disruptive element that helps bring new players and new ideas into the military’s traditional research, development and acquisition (RD&A) process. They create new avenues for the PLA to directly interface with younger innovators from within and outside of the PLA. The added modern vibes to these “games” and “challenges” also help increase the appeal of the PLA’s innovation culture. In the long run, a more open and modern innovation culture will no doubt be conducive to its recruitment of a college-educated, tech-savvy younger workforce.

The 2020 “Stratagem at Heart, Jointness to Win (谋略方寸 联合制胜)” Joint Operations Challenge

This is perhaps one of the most interesting, and also the newest, AI competition series that the PLA is currently running. As the name of the competition suggests, participants are invited to compete in designing the best algorithms for joint operations. The official website of the contest lists the Central Military Commission (CMC) Equipment Development Department (EDD) as the host of the AI contest, with the support of the China Electronics Technology Group Corporation (CETC), National University of Defense Technology (NUDT) and China Aerospace Science and Industry Corporation (CASIC). In a widely circulated poster for the game, CETC’s 28th Research Institute (RI) is listed as the actual CETC entity involved.

Most noteworthy in this game is that the competing algorithms are required to operate in “joint island strike” operations to resolve a “disputed sovereignty issue pertaining to an island that is currently occupied by the adversary.” This is essentially a euphemism for a Taiwan invasion scenario. The Red Team (Blue in the U.S. context) is required to use the simulated operations platform created by the host of the game to “develop and train AI” to carry out decision-making and operation planning for complex operations that include “target reconnaissance, electromagnetic countermeasures and coordinated firepower strikes.” According to the game’s design, Red’s intention is to “strike two island command posts so as to paralyze its command and control systems,” and both Red and Blue (Red in the U.S. context) teams are allowed to operate “close to a hundred combat units comprising 15 categories, such as early warning aircraft, fighters, bombers, electronic warfare aircraft, ground radars, air defense missiles, anti-ship missiles and destroyers.”

PLA Rocket Force (PLARF) and its “Intelligent Rocket and Fire Eyes” AI Challenge

The PLARF’s “Intelligent Rocket and Fire Eyes” (智箭火眼) AI Challenge mainly focuses on utilizing AI to enhance precision strike capabilities. Hosted by the PLARF Equipment Department, with support from the PLARF Research Academy and CASIC Fourth Academy’s 17th Research Institute (RI), it is scheduled to run from August to December 2020. The 17th RI, established in 1968, is responsible for solid fueled ballistic missile guidance, navigation, and control systems. Also known as the Beijing Institute of Control and Electronic Technology, the 17th RI likely has the lead for systems integration work on the anti-ship ballistic missile (ASBM) guidance, navigation, and control sub-system, including software development. In 2007, it opened a C4ISR National Defense Key Laboratory.

The competition is designed to focus on five key subject areas: image recognition against typical background; image recognition against complex background; multimodal remote sensing image registration capability; remote sensing imagery recognition capability; and imagery detection and recognition capability under complex conditions (by invitation only).

The intense focus on applying AI to imagery recognition, detection and registration, combined with the direct involvement of the 17th RI in the PLARF AI Challenge, confirms that the PLA in general, and the PLARF in particular, have a clear intention of advancing its precision-strike capabilities with AI. Although the result won’t be known until the end of this year, and outside observers may never be able to know the key data referenced throughout the game, it is still worth tracking and understanding the winning team’s achievement. The winning team members’ affiliations and/or future employment within the PLA’s defense engineering enterprises may also serve as useful references for understanding its future technological focuses.

PLA Strategic Support Force (PLASSF) and Its “Intelligent Space Cup (天智杯)”

The PLASSF’s AI Challenge was created in 2018. In terms of the organizational structure, it is slightly different from the previous two AI challenges discussed above. It is hosted by the PLASSF Space Systems Department’s Equipment Department, with direct guidance from the CMC EDD. Most of the known units subordinate to the PLASSF Space Systems Department and key elements of the EDD were previously under the General Armaments Department prior to the 2016 military reform. The PLASSF AI Challenge reflects the close working relationship the PLASSF’s military space elements and the EDD continue to share.

The 2019 competition, which lasted around six months, was concluded in December of that year. It had two key areas of focus: “intelligent detection and recognition of remote sensing imagery” and “applications of intelligent processing for survey geography and meteorological and hydrological data.”

The second “Intelligent Space Cup” competition was scheduled to take place in late 2020, but it may have been delayed due to the COVID-19 pandemic. It kept the “meteorological and hydrological data processing” component from the 2019 game as one of its key areas of focus, and added: “application of remote sensing data interpretation” and “intelligent processing applications for situation awareness data.”

The PLASSF’s AI Challenge has a clear emphasis on applying AI to data processing and analytics. This focus aligns well with the perceived mission of the PLASSF Space Systems Department’s intelligence support function. Considering that official information about the PLASSF’s actual peacetime and wartime functions remain scarce, the focus areas of the game highlight and confirms the importance the SSF likely places on taking full advantage of the imagery as well as the meteorological and hydrological data fetched by its remote-sensing satellites.

PLA Air Force (PLAAF) and Its “Intelligent Aerospace (智胜空天)” and “Unmanned Dominance (无人争锋)” Competitions

The PLAAF has been actively involved in running at least two series of competitions that both put an emphasis on unmanned aerial vehicle (UAV) technologies. The “Intelligent Aerospace UAV Challenge” features UAV designs and capabilities, whereas the “Unmanned Dominance” challenge invites its participants to develop “drone swarm” technologies.

Hosted by the PLAAF Engineering University (PEU), both the 2018 and 2019 “Intelligent Aerospace UAV Challenge” contests lasted about three days. The competition in 2018 attracted about 100 competing teams whereas the 2019 contest had 112 teams participating. Participants hail from a wide array of backgrounds, ranging from military and civilian entities, including NUDT, Xiamen University and the Chinese Flight Test Establishment in Yanliang. The 2018 contest took place on the campus of the Air and Missile Defense Academy of the PEU. The contest includes UAV multi-target recognition, target strikes, real-time planning and autonomous collision avoidance and 2 vs 2 UAV force-on-force air combat.

The first “Unmanned Dominance Drone Swarm Challenge” was also created in 2018, which took place in CETC’s industry park in Laishui county of Baoding, Hebei. The second contest was held in Xi’an, Shaanxi. This series of competitions is managed by the PLAAF Research Academy and the CETC Academy of Electronic Science, with support from the Chinese Aeronautical Establishment (CAE), Beijing Institute of Technology (BIT), Tsinghua University and other institutions. One notable feature of this contest series is that it is closely linked to the PLAAF engineering development. The winning technologies are said to be given priority consideration in the PLAAF’s equipment preliminary research.

The “drone swarm” contest also created a dedicated WeChat account (“Intelligent Competition/智能争锋”) for information sharing. During the interval of the competitions, however, the account has populated its posts with articles and commentaries that track almost every move made by the U.S. military with regard to AI applications. In October 2020, this account posted an article titled, “CETC has created China’s first operational ‘Drone Swarm’,” along with a video demonstrating that CETC’s Academy of Electronic Science has recently completed a flight test of its ground-air coordinated fixed wing UAV swarm systems, which carried out ground reconnaissance-strike and precision strikes. It may be inferred that at least part of the “operational drone swarm” systems resembled achievement from the “Unmanned Dominance” contest, although a more conclusive assessment of the linkage will not be reached until more official information is made available.

PLA Army (PLAA) “Unmanned Platform” Competition and PLA Navy (PLAN) “UUV Tech Demo Exchange Activities”

The PLAA appears to have the longest history of such AI-themed competitions. Notably, the idea behind the Army Contest can be traced back to GAD’s 2014 unmanned platform competition. According to a PLA officer from the GAD Army Equipment Office’s S&T Research and Acquisition Department, the competition was created to “verify autonomous vehicles’ battlefield adaptability” as well as to “test out new avenues for military innovation and accelerate the development of autonomous technologies.” Although its 2020 contest has been postponed until 2021, there are a number of categories of the contest that are worth monitoring, such as the ground-air “recon-strike” unmanned swarm contest and the bio-inspired robots that fight alongside the PLA human warfighters.

Since it was created in 2016, the PLAA contests have drawn interest from military and civilian participants alike. Similar to other service-themed AI challenges, the PLAA Equipment Department works through its Research Academy to set up the requirements and content for the competition, which is open to institutions from both within and outside of the PLA. The Army Challenge 2018 took place on the test site of the Army Research Academy’s Institute of Armored Forces, and its participants were invited to compete for designing autonomous vehicles to carry out autonomous mobility, reconnaissance, “air and ground coordinated blockade and control” as well as bio-robotics contest and other high-mobility autonomous mountain vehicle transportation, reconnaissance, strikes and penetrations. The competition is often followed by a high-level summit focusing on the same themes as the contests. Both competition and summit serve as a platform for information sharing among operational units, academic institutions, the defense industry and S&T research institutions.

Finally, the PLAN appears to be a latecomer in creating such competition programs. It was not until May 2020 that the EDD WEAIN website posted the information about the first UUV Tech Demo Exchange Event scheduled to take place in September 2020 in Kunming, Yunnan. More details of the game are pending, except that the “exchange,” according to the official description, is composed of three parts: actual small UUV (under 1 ton) contest, an exhibition and a featured conference on UUV technologies.

Conclusion

Key PLA RD&A organizations have taken new measures to accelerate military innovation. The EDD-led, competition-centric practices for integrating emerging AI technologies demonstrate the PLA’s seriousness in strengthening its support for military innovation. The significance of what the PLA is doing lies not in the technologies on display per se; rather, it is the changing mindset driving the PLA’s new activities that is of critical importance. The PLA has taken on significant effort to reform its culture for innovation, something that should be taken seriously by the U.S. defense community and China strategists who are serious about out-thinking and out-innovating the PLA.

Dr. Marcus Clay is an analyst with the U.S. Air Force’s China Aerospace Studies Institute (CASI). The views expressed are those of the author and do not reflect the official policy or position of the U.S. Air Force, Department of Defense, or the U.S. government.