Under the Ministry of Culture and Tourism of the People’s Republic of China
Sponsored by the China Academy of Art
2025 No. 1 Issue, Total No. 254
Core Journal of Chinese Humanities and Social Sciences Journal AMI Comprehensive Evaluation (A Journal)
Source Journal of the Chinese Social Sciences Citation Index (CSSCI) Extended Edition
RCCSE Chinese Core Academic Journal (A)
Important Reprint Source Journal of Renmin University of China “Copying Newspaper and Periodical Materials”
Special Planning
AI Art Creation Based on Agents——Taking “Symbiosis of Agents” as an Example
【Abstract】“Symbiosis of Agents” debuted at the 14th National Art Exhibition, highlighting the cross-border integration of artificial intelligence and artistic creation. “Symbiosis of Agents” employs the cutting-edge technology paradigm of agents, which, through direct interaction with the environment, can exhibit human-like perception, planning, and decision-making abilities, paving new avenues for the creative fusion of art and AI. The work also delves into the consciousness and intentions of agents, enriching the understanding of the Turing problem and showcasing the potential applications of multi-agent systems in the real world, as well as their roles in artistic expression and social commentary.
【Keywords】Artificial Intelligence; Agents; Tech Art; Turing Problem
In the current research field of the cross-border integration of artificial intelligence (AI) and artistic creation, my work “Symbiosis of Agents” selected the cutting-edge paradigm of AI agents as the main creative object. Agents utilize data more efficiently than traditional data-driven AI systems. Currently, AI research trends are shifting from the data-dependent “Internet AI” to an “agent” model centered on interaction with the environment, which will open new pathways for the integration of art and AI creation. An agent refers to an AI system with a physical or virtual body, encompassing a wide range of definitions, from humanoid robots in future industries to various entities in video games, including non-player characters (NPCs), often characterized by generality and embodiment, with research fields spanning robotics and biomimetics. Researchers simulate physical environments and utilize advanced computational capabilities to design and train intelligent systems that can perceive the environment and act, aiming to transfer their intelligence to the real world for autonomous decision-making and interactive task completion.
▼Figure 1 Chen Baoyang “Symbiosis of Agents” Multi-Agent Device (Mechanical Arm, Robot Dog, Lighting, Smoke, Screen) 6m×6m×2m 2024▼
1. From Generative Tasks to Agent Systems
Looking at the development of the integration of AI and art, there are currently three main forms of creation—generative tasks, AI-driven nonlinear human-computer interaction, and multi-agent constructed ecosystems. These three not only showcase the complementarity of technology and art but also reflect the deep evolutionary paths of artistic creation. The development of these forms stems not only from technological advances but also signifies the evolution of creators’ thoughts on the Turing problem, shifting from the initial “creation and imagination of machines” to deeper explorations of “machine consciousness and intention.” Generative tasks, as a more mature path in AI art development, focus on the self-executing capabilities of algorithms, autonomously generating images, videos, and three-dimensional works. Through data-driven learning and analytical capabilities, machines can independently complete creative tasks, reflecting the basic ability of machines to mimic the creative processes of human artists. This rule-based automated process initially responds to the challenges of the Turing test, demonstrating the potential of machines in processing and recreating visual arts. With the advancement of new technologies, AI-driven nonlinear human-computer interaction has emerged, which not only increases the interactivity of artworks but also transforms the artistic creation process into a dynamic system involving both humans and machines. At this stage, AI is not just a tool for executing preset tasks but can also adjust content in real time based on audience behavior and reactions, thus introducing a certain degree of “intention” and “feedback response” into artistic expression, making the issue of machine “consciousness” a possible discussion point. The complex ecosystems constructed by multi-agents elevate AI art to new heights, where agents not only interact to create complex ecological landscapes but also exhibit adaptive and learning capabilities akin to organisms. In this model, interactions between agents may produce new behavioral patterns that were not pre-set, reflecting a higher level of systemic “intention,” providing richer material and philosophical reflections for exploring machine consciousness.
The current research focus in AI art creation is on generative tasks, involving the use of various tools to generate images, videos, and reconstruct three-dimensional worlds. Existing research views this type of generative creation as a continuation of digital visualization (pixel art, character art, glitch art, etc.). The following paradigms have emerged in AI art creation: exploring artistic innovation through programming and computer-aided technologies, integrating geometric patterns and mathematical principles, emphasizing the randomness and systematic nature of the creative process. The core of this paradigm is to utilize predefined algorithms or rules to drive the creative process, which unfolds based on the complex relationships between rules and freedom, predictability and uncertainty.
Currently, research on generative tasks mainly focuses on data-driven “Internet AI,” commonly referred to as generative artificial intelligence, a mode of operation that analyzes and learns from a vast amount of art data on the internet to automatically or assistively generate new artworks. Relevant studies also involve issues such as authorship transfer, copyright, and data ethics. It is noteworthy that these AI systems create images by analyzing different probability distributions, which significantly differs from the paths of “observation, thinking, and hands-on” in traditional artistic creation processes. Therefore, research addressing this difference is particularly important, and the perspective of agents provides a valuable lens for this, as agents can simulate the dynamic decision-making processes of human artists during creation and achieve deeper levels of interactivity and adaptability, thereby more comprehensively understanding and expressing the complexities of art.
In the realm of interactive art, the core viewpoint holds that AI can innovate traditional linear interaction models by introducing dynamic interaction mechanisms based on real-time computation, making the connection between the audience and the artwork exhibit more vivid nonlinear characteristics. As a result, the work can respond instantly to audience reactions, achieving personalized displays, significantly enhancing audience engagement and interaction quality, further highlighting the potential of AI in enhancing interactivity and practicality in artistic creation.
Existing research summarizes the creative paradigms from the perspective of human-computer interaction, utilizing sensory transmission, behavioral interaction, active and passive exploration, and nonlinear assembly to interact with the audience. First, artists purposefully manipulate multiple sensory channels of the audience, such as visual, auditory, and olfactory, through integrated hardware and software systems, to achieve deep psychological communication. Second, artworks convey artistic content by triggering spontaneous behaviors or conditioned reflexes in the audience. Third, audiences are guided to engage in autonomous or passive exploration in programmed environments, promoting independent thinking and enhancing their sense of presence. Finally, artists set multiple interaction nodes to encourage audiences to nonlinearly reorganize information, thus creating a polysemous cognitive experience. These methods collectively expand the dimensions of interaction between art and the audience.
As the application of agent technology increases, the intelligent “artificial systems” constructed by artists have expanded to include complex interactions between agents. Preliminary theoretical research has primarily been limited to using Martin Heidegger’s concepts of “worlding” and “dasein.” In the application of multi-agent systems (MAS), agents not only embody the dynamic existence and manifestation of the world described by Heidegger but also demonstrate system interactivity through cooperation and competition among agents, as well as their adaptability and learning capabilities. This research essentially applies and showcases Heidegger’s philosophical thoughts on the continuous generation of the world, though its singular research perspective also exposes a neglect of other possible interaction patterns and theoretical frameworks.
In the aforementioned field of AI and art integration, artistic practice has evolved from singular AI generative tasks to interactive and collaborative multi-agent work, demonstrating a transition from technical automation to complex systemic interactions. This shift not only deepens the exploration of the Turing problem—from initial explorations of AI creativity to in-depth studies of machine consciousness and intention—but also provides rich philosophical and technical backgrounds for my artistic creation. Within this framework, the creation focuses on dynamic collaboration and competition within multi-agent systems, exploring how these interactions construct complex social simulation landscapes. By employing embodied AI, the work not only investigates the performance and environmental perception of agents in the physical world but also showcases a dynamic, self-evolving artificial ecosystem through the behaviors of these agents. This ecosystem utilizes AI to generate visual images, audio, and three-dimensional objects, while also realizing Heidegger’s concept of “worlding,” creating interconnections and meanings among objects through the interactions of agents. In my agent-based creation, each element within the work is an “agent,” capable of perceiving the environment, making decisions, and executing actions. Each agent has its own goals, which can be cooperative, competitive, or a combination of both. They can independently make decisions and interact to achieve collective or individual goals. Each agent has its own strategies, knowledge, and goal-oriented persona, while the overall behavior of the system is the result of interactions among these agents.
In exploring the integration of AI and art from the perspective of agents, I gradually delved into the question of whether agents possess consciousness through two preceding works—”Congratulations, but I am Next to Him” (2019) and “Chasing the Mist” (2020). These two works reflect the progression from initial agent autonomy to more complex interactive systems, laying the groundwork for understanding and developing more advanced agent systems. “Congratulations, but I am Next to Him” explores the behavioral patterns of machines in rule-driven environments through an agent car that traverses a maze in a cyclical manner. This work symbolically represents the interaction between humans and algorithms in modern life through the actions of the agent car in the AI-generated continuous maze. Each traversal by the agent is not only a pursuit of a goal but also reflects a deeper philosophical question: Is the agent merely an executor of programmed logic, or can it embody some form of preliminary consciousness and decision-making ability? Subsequently, “Chasing the Mist” expands the understanding of social behaviors of agents by constructing an interactive scene where two embodied agents chase each other. In this work, the interactions and fundamental interactions between agents not only showcase the dynamics of multi-agent systems but also complicate the exploration of agent consciousness. The audience influences agent behavior by setting virtual obstacles, effectively building part of this ecosystem, and this interactive experience deepens the understanding of agent autonomy and collective behavior. This is not just an observation of agent behavior but also a philosophical exploration of its decision-making processes, reflecting the role and responsibility of humans in constructing and controlling complex systems. Through the progressive development of these two works, I explored the application of agents in artistic creation and gradually delved into the philosophical and technical discussions of their potential consciousness, providing a rich theoretical and practical foundation for the subsequent work “Symbiosis of Agents” to conduct in-depth research on the interactions and cooperation of agents within multi-agent systems.
▼Figure 2 Chen Baoyang “Chasing the Mist” Interactive Device Size Variable 2020▼
“Symbiosis of Agents” utilizes commonly seen mechanical arms, reconnaissance quadruped robots, lighting, smoke, and screens in industrial production to construct a narrative environment exploring the issues of human interstellar migration and AI symbiosis. Through polarizing film, the pristine screen displays people’s aspirations for a beautiful natural environment in the future, while robots walking on the broken mirror suggest the construction and collapse between natural ecology and digital technology. Additionally, I created a virtual experimental field named “Phantom Mirage” for the interactions of various agents in the work. This experimental field integrates narrative, multi-agents, and cognitive science, studying the impact of new technologies on society through the competition and cooperation among agents, particularly examining how technology will influence human and social structures. In this virtual ecological environment, agents are not only part of the ecosystem but also the main force shaping and interpreting this environment. In the “Phantom Mirage” system, software, robots, and humans can work together to simulate complex real-life scenarios. The artificial ecosystem constructed using these technologies not only demonstrates the survival and evolutionary states of multi-agent systems but also attempts to present a complete, dynamic, and self-evolving ecological environment. This is not merely about creating captivating digital landscapes but also about exploring and showcasing a vibrant, continuously developing ecosystem.
2. Rules, Narratives, and Multi-Agent Simulation
In a self-evolving multi-agent system, each agent must follow certain rules or strategies to interact with other agents. In AI reinforcement learning environments, these rules or strategies are typically determined based on reward and punishment mechanisms. Constructing artificial ecosystems shares many similarities with creating multi-agent environments. Therefore, it is necessary to clarify how participants in this system or ecology should act and what consequences their actions will bring. These behavioral norms and consequences are essentially the rules of this system or ecology.
The core of “Symbiosis of Agents” is to construct a set of rules within the framework of “Phantom Mirage,” which includes a narrative structure and specific execution strategies. I prefer to refer to this work as a simulation within the “Phantom Mirage” simulation system, envisioning a scenario of AI and humans coexisting in constructing ecosystems on Mars. The narrative revolves around environmental transformation during human deep-space exploration, metaphorically “planting” a tree to signify humanity’s beautiful expectations of altering Mars’ atmosphere to make it habitable. In the near future, humans will invent and upgrade recyclable rockets through continuous attempts, allowing those developing Mars to no longer be limited to one-way tickets. Thus, I set various AI-driven machines as the vanguard in “Phantom Mirage,” heading to Mars to build the infrastructure required for human habitation and modify the Martian climate in preparation for humanity’s arrival.
This work consists of multiple agents and non-agents, aiming to achieve the sustainable and healthy growth of trees through highly coordinated interactions. The core of the system is two groups of plants, which serve as non-agents monitored and maintained by agents such as mechanical arms, robot dogs, lighting, smoke machines, and weather prediction systems. The mechanical arm adjusts the position of the plants and is associated with lighting and smoke machines to ensure optimal light and humidity conditions for the plants. The robot dog takes on reconnaissance and security duties, assisting the mechanical arm in caring for the plants and monitoring the environment. The lighting agent automatically adjusts the direction and intensity of light as needed, while the smoke machine adjusts humidity supply based on feedback from the environment and the mechanical arm. The weather prediction system provides forecasts for future weather conditions, helping other agents optimize plant care strategies. All interactions among agents are optimized and configured around the plants’ health growth needs to achieve the system’s comprehensive functions and goals. Additionally, this system interacts with the audience through the robot dog’s perspective, which alters its decisions based on the collective behavior of the audience, thus influencing and adjusting the collaborative state among agents.
▼Figure3 Chen Baoyang “Symbiosis of Agents” Multi-Agent Device 6m×6m×2m 2024▼
In the “Phantom Mirage” virtual simulation system, these rules not only support effective interactions among agents but also ensure that the entire system can simulate the complexity and dynamics of the real world. Through a carefully designed rule system, agents can self-evolve and adapt in the “Phantom Mirage” virtual environment, demonstrating behavior patterns similar to those in the real world. The establishment and implementation of such rules essentially simulate the process by which organisms in natural ecosystems adjust their behaviors based on environmental feedback. Therefore, this not only showcases the application of technology but also expresses a profound exploration and reflection on the management and control theories of complex systems to ensure coordination among agents and the overall health of the system. Such attempts not only explore the impact of technology on social structures but also simulate a possible future: how intelligent systems and human society will influence and adapt to each other.
The aforementioned works collectively reveal that multi-agent systems and embodied intelligence can exhibit complex behaviors based on simple rules that each agent follows, providing rich expressive means for artistic creation. In fact, this idea has historical precedents in the history of art. In the early integration of AI and art, it can even be traced back to the rise of generative art, where artists set rules and relied on human or machine to deduce these rules, which provided an artistic historical anchor for the creation of rules in multi-agent systems. In the 1950s and 1960s, artists began to create artworks using mathematical and logical rules. The core of this art form lies in using predefined algorithms or rules to drive the artistic creation process. With the advancement of computer technology, this practice has been widely applied in the field of digital art. Artists like Harold Cohen and Manfred Mohr are pioneers in this field, creating art through programming and computer-aided drawing techniques. In the art movements of the 20th century, artists often deduced and created works based on preset rules and structures, such as Bridget Riley and Victor Vasarely, who constructed visual illusion works using strict geometric patterns. In fractal art, artists use mathematical principles to produce infinite detail and complexity, which is based on fractal theory. Similarly, conceptual artist Sol LeWitt created works based on specific instructions or rules, with his “Wall Drawing” series executed step by step according to explicit instructions, resulting in different outcomes due to variations in the execution process. These historical examples show that artists have long adopted systematic approaches to control and guide creation through defining rules and processes, providing a profound artistic historical foundation for rule creation in multi-agent systems.
When examining the similarities and differences between the rules set in multi-agent systems and those in generative art tasks, it should first be noted that there are significant differences in the purposes of rule setting between the two. Generative art relies on algorithms and mathematical principles for automated creation, primarily through data-driven methods, utilizing large-scale art data to generate new visual representations, emphasizing randomness and systematicity, thus broadening the boundaries of artistic expression. In contrast, the rules in multi-agent systems are mainly used to guide agents on how to interact effectively with other entities, usually based on reward and punishment mechanisms to optimize the overall performance of the system or provide pathways to solve specific problems; thus, their rules emphasize clarity and consistency. In terms of execution, AI art creation automatically processes and generates artworks through complex computational models, while agents in multi-agent systems adjust their behaviors based on dynamic environmental feedback. Regarding the predictability of outcomes, generative art has a certain level of uncertainty due to its reliance on probability distributions, while multi-agent systems ensure expected system behaviors through precisely designed rules. Therefore, although both utilize predefined rules to drive processes, they present clear differences in purpose, execution methods, and outcome predictability, reflecting the diversity of rules in modern technological applications and the ongoing exploration of balancing control with freedom, predictability with uncertainty.
In the aforementioned works, I no longer merely assume the role of a traditional creator but more as a designer or director, setting rules for the system and then objectively observing, interpreting, and even intervening in the system’s performance. In multi-agent works, each agent can be viewed as a narrative unit, acting according to established basic rules, gradually weaving complex behaviors and storylines. This approach not only showcases the potential artistic value of multi-agent systems but also emphasizes the core position of narrative in art. I am not just a storyteller but also an observer and interpreter, setting rules for this narrative system and extracting and constructing stories from it. Each installation in this work can be seen as an open narrative experiment, demonstrating the infinite possibilities of narrative dynamics in artistic creation.
3. The Simulation Interface of Multi-Agents
By exploring the setting and application of rules in multi-agent systems, the related concepts provide a perfect transition to introduce the concept of the “Phantom Mirage” simulation system. “Phantom Mirage” serves as a product combining art and technology, reflecting not only the advanced nature of digital simulation technology but also profoundly revealing the boundaries between reality and the virtual. Through such a system, we can further explore and validate phenomena that are difficult to observe or simulate in the real world, thereby fostering new understanding and innovation through the combination of art and technology.
The choice of the name “Phantom Mirage” provides a deep symbolic meaning for the simulation system, reflecting its dual properties between reality and the virtual. Literally, “phantom” usually refers to something unreal or simulated, while “mirage” is a phenomenon in nature that creates an illusion, often appearing on the distant horizon, distorting or inverting distant objects. The combination of these two terms weaves a dual illusion that is neither purely actual nor entirely fictional but wanders in the foggy realm between reality and dreams. Simulation is essentially a technology that attempts to replicate phenomena or system behaviors of the real world. By creating such a system, the work achieves the observation and analysis of phenomena that may not be directly observable in the real world within a controlled environment. Thus, simulation is both real—based on data and laws of the real world—and false—it remains a controlled, designed model that cannot capture all the details and complexities of the real world. I named this system “Phantom Mirage” to emphasize that it shows one side of reality, helping us understand and predict the future through precise simulation and prediction, while also exhibiting an illusory side, as it can never achieve the infinite complexity and unpredictability of the real world. Additionally, this name suggests the inherent limitations and potential misleading nature of such simulation—just like a mirage, it appears real but ultimately is an illusion. “Phantom Mirage” reminds us that while simulation can provide valuable insights and predictions, we must also recognize its essential incompleteness and possible biases. This dual negation’s philosophical depth helps me maintain caution and introspection when utilizing these powerful tools, ensuring that I am not deceived by their appearances and can fully explore the subtle rifts between reality and the virtual. Moreover, through artistic exploration of the boundaries between the virtual and the real world, we can observe how technology shapes human perceptions of reality and emotional experiences, providing unique perspectives for deeply understanding the aforementioned boundaries.
▼Figure4 Chen Baoyang “Symbiosis of Agents” AI System Interface▼
I create a layer of separation in the digital world, an interface. Through this interface, one can see the Antarctic glaciers melting but cannot hear the thunderous collapse of the glaciers. After caring for, constructing, and examining the digital world, I deeply reflect on the ambiguity of my position in the overlapping real and virtual worlds, thus conceptualizing that special tree in “Symbiosis of Agents.” It may represent me or anyone, appearing in a romantic yet sorrowful scene. In “Symbiosis of Agents,” the tree no longer takes root in the soil but is controlled by a mechanical arm. In this AI-driven environment, prioritizing the health and growth needs of the tree overlooks its actual requirements. I deliberately removed rules that prevent harm or adverse effects on the tree and intentionally set “incorrect” parameters—the needs and acquisition frequency of sunlight and moisture for the tree are artificially defined. This causes the tree to endure immense pressure when the mechanical arm raises or lowers it too frequently. The mechanical arm represents the inherent logic of the digital world, which may subtly disrupt the foundation of what it means to be human. Until the end, when the branches break and leaves scatter, the mist gradually dissipates, but the mechanical arm continues to operate as usual, while that tree has vanished. Such a narrative scene showcases the complex and subtle relationship between humanity and the technological world it creates.
5 Chen Baoyang “Symbiosis of Agents” Multi-Agent Device 6m×6m×2m 2024
Although AI simulation systems can showcase different future directions, in the real world, we must make relatively singular choices based on current information and environments, attempting to move forward along determined paths. In “Symbiosis of Agents,” optical devices explore the possibilities of multiple realities and choices. In a series of carefully designed optical devices, the close connection between the infinite possibilities of multi-agent simulation and real choices is cleverly demonstrated. The optical devices in the work have two viewing surfaces, each displaying unique “slices” from the “Phantom Mirage” system at different angles, symbolizing the states of the multi-agent system in simulating complex environments, such as the diversified future scenarios revealed under the premise of Mars colonization. Through this visual and conceptual presentation, the audience is guided to reflect on how multi-agent simulation theoretically provides a platform to explore multiple possibilities and reconsider how the decisions we make in real life influence and shape our future. This work not only examines the moral, social, and cultural challenges that may arise in the coexistence of humans and artificial intelligence but also emphasizes the importance of the choices made by humans when facing significant tasks like Mars colonization.
“Symbiosis of Agents” borrows multi-agent simulation technology to delve into various complex issues within the field of artificial intelligence, set against the backdrop of human and AI coexistence on Mars. By viewing artificial intelligence as “the true residents of Mars,” I attempt to discuss the profound ethical and responsibility issues arising from humanity’s use of it, exploring the resident identity of artificial intelligence in Mars colonization and its social ethics and philosophical implications. This is not only a discussion about technology but also a profound reflection on marginalized and overlooked groups in human history. I hope that visitors can reconsider the role, contributions, and historical significance of artificial intelligence in future societies through their observations of the exhibits.
This multi-agent simulation system’s construction aims not only to predict the future but also to shape the entire project into a reverse archaeological exploration landscape through in-depth exploration and reassessment of human historical actions. By imagining the future to reinterpret and reflect on the past, this method helps us understand the long-term consequences of historical actions and prompts us to reevaluate the significance and impact of those actions from a future perspective. In setting the first law of simulation—the basic behavioral rules of agents—”Symbiosis of Agents” emphasizes the importance of cautious analysis and choice regarding simulation results, reminding viewers of the risks of over-reliance on machines for decision-making. While AI can process complex data and predict potential problems, they lack human moral intuition and cultural sensitivity, making human involvement indispensable in interpreting results and making moral judgments: recognizing and understanding the complex interactions among agents and their cumulative effects to predict potential future problems or challenges, a process that should not solely rely on machine-generated data and analysis results. Although current AI has not yet achieved true self-awareness, as technology develops, we must consider the ethical and philosophical issues surrounding the potential for intelligent systems to possess higher forms of consciousness in the future, including rights, responsibilities, and their roles in society. For instance, if machines can autonomously make moral decisions, how can we ensure their decisions align with human values? In the context of future archaeology, using imagination about the future to reflect on the past, we must handle and utilize technology cautiously to ensure that technological development aligns with humanity’s long-term interests, without neglecting potential risks and moral dilemmas.
“Symbiosis of Agents” showcases the exploration and application of cutting-edge technologies in artificial intelligence, simulating a dynamically changing social simulation ecological system through the establishment of complex interaction rules among agents and exploring the relationships between humans and technology, nature and the digital world. These agents act within the set rule framework, as if performing in a meticulously designed theater, where each interaction is a recreation of existing narratives.
In my creative process, the integration of AI and art not only reflects the interaction between technology and art but also profoundly practices the core concepts of humanistic spirit and critical reflection. The creative process emphasizes the isomorphism of artistic and technological experimental exploration, focusing on the impact of technological development on human mental states and social psychology, while consistently examining technological progress from a critical perspective. Furthermore, through the lateral and creative thinking of art, this creation not only utilizes existing technologies but also promotes reflections on scientific innovation through the interaction of art and technology, thus demonstrating the enlightening and supportive role of artistic thinking in technological innovation.
I will continue to deepen this interdisciplinary artistic practice, not only applying agent technology to excavate and express profound understandings of human culture but also introducing fundamental biological rules to provide new guiding directions for agent system behavior settings. This will enable my artworks to not only showcase technology but also become a symphony of culture and science. In this way, the combination of ancient myths and modern science will bring new understanding and cognitive perspectives. In future works, I will further explore how these agents achieve self-adjustment and evolution in simulated environments and how they reflect humanity’s understanding of ecosystems and biodiversity. By integrating biological principles into the construction and simulation of agents, new works will delve deeper into how agents simulate complex life processes in artistic creation and how these processes will influence human perceptions of the environment and society.
(For reading convenience, footnotes omitted.)
Chen Baoyang: Lecturer at the School of Experimental Art, Central Academy of Fine Arts