Throughout history, people have been interacting with the world in new ways with the help of technology to gain a more comprehensive understanding of the world and ourselves.
Past experience has made us deeply aware that, whether it is from academia or industry, it is quite difficult to use current methods to achieve general artificial intelligence (AGI) in a short period of time. Machines not only need to learn massive amounts of data from all walks of life, but also need new algorithms and other ways to better understand human knowledge itself.
In the digital entertainment field dominated by games, we see the application scenarios and opportunities of artificial intelligence. The game itself has highly structured data, which can provide artificial intelligence with many limited scenarios for training.
During the game experience, players will also generate a large amount of behavioral data when interacting with digital content, which provides an indispensable motivation for the further iteration and optimization of artificial intelligence.
From the live-action running games represented by Dungeons and Dragons in the 1960s to the open world games represented by GTA 5 and Red Dead Redemption 2, many gamers and developers have always had one Dream: To be able to realize the infinite possibilities of free interaction in Dungeons and Dragons, as well as dynamic personalized plot and narrative experience in video games.
Of course, we are no exception. In the two years since our establishment, we have continuously absorbed opinions and feedback from game companies and global game markets around the world, accumulated technology and iterated, and gradually explored a set of methods and products, which can be used in games Real-time and dynamically generate personalized interactive content with players.
We named this core technology Chaos Box, which is an AI narrative engine based on deep reinforcement learning. Without any scripting, it can control the behavioral logic of virtual characters in the game, and Let it spontaneously produce very intelligent behavior.
Chaos Box not only realizes a more personalized and dynamic plot exploration, but also intelligently generates scripts and multi-line plots for fixed goals or endings. It can also generate intelligent NPCs, intelligent organisms and intelligent environments with different behavior patterns and dialogues on a large scale in the game.
In the traditional game production process, people need to use decision trees and state machine technology to clearly define and design the objects, behaviors and results in each event in order to achieve the diversity of content and Complexity.
In order to better attract players, the game needs to be constantly updated. If people want to add new items, interaction methods, event situations, etc. to the existing decision tree, they must modify and adjust all related decision trees.
However, developers spend a lot of manpower and time to create a small amount of content, and users will consume it at an alarming rate in a short time.
Different from the decision tree, Chaos Box provides the content in the game with a “brain” capable of autonomous decision-making, intelligently generating and controlling the game content itself and the interaction logic with the player.
With simple rules and settings, game developers can quickly generate dynamic and personalized plots and events, and digital content such as NPCs that can interact with players intelligently and behaviorally.
At the same time, Chaos Box can be arbitrarily combined with traditional decision trees. Game developers can choose to replace some nodes of the existing decision tree with Chaos Box one by one to create unlimited content interaction possibilities in local scenes; or they can use Chaos Box to cover all game scenes, thus making a completely open and dynamic content that can be automatically generated.
In order to allow content creators and developers to use Chaos Box to intelligently generate dynamic digital content in a simple and efficient way, we have integrated Chaos Box into a cloud-native platform: Morpheus Cloud.
Morpheus Cloud is a cloud native service based on Chaos Box algorithm. When the game is connected to the Chaos Box algorithm, Morpheus Cloud can generate intelligent dynamic content in real time as the player interacts with the game content.
The way the player interacts with the game includes but is not limited to text, voice, and all interactive behaviors that are allowed in the game. According to these different interactions, all virtual characters, creatures, and objects in the environment in the game can generate different content in real time.
When game developers use Morpheus Cloud for development, they only need to set very simple objects that players can interact with in a certain scene, such as characters, objects, behaviors, etc., and combine all interaction relationships It is clear that training can be carried out with the help of Chaos Box algorithm, so that interactive content such as dynamic and personalized plots and intelligent NPCs can be realized in this scene.
In fact, cutting-edge game technology has brought new possibilities to the industry. Morpheus Cloud and AI can help us achieve more effects and experiences that were impossible before.
In the game, Morpheus Cloud can not only generate personalized intelligent plots, narratives, NPCs and other interactive content simply, automatically, and efficiently, but also generate motion animations and effects in real time through dynamic prediction of character movements.
In order to make content generation more intelligent and effective, Morpheus Cloud can also help developers convert text and natural language into real-time animation. It can convert scripts, novels, news and other text creatives into visual images and animation effects at a very fast speed.
In addition, Morpheus Cloud can dynamically generate different styles of rhythmic music and levels according to the characteristics of different games and the interaction of players.
At the same time, in the process of continuous automated testing of games, Morpheus Cloud can also continuously help game developers optimize operational strategies based on data analysis of the interaction between players and games. For example: automatically balance the economic system in the game, intelligently generate various types of game activities, etc.
With the continuous development of the virtual world, games will exist as the main form of interaction between users and digital content; and gamified content interaction will gradually penetrate into other fields. As a result, while changing the way of interaction between users and the pan-content field, it also brings new needs of users for massive and personalized real-time content interaction.
In addition to bringing efficient and intelligent content generation in games, Morpheus Cloud can also bring new possibilities in other fields such as media, social networking, live broadcasting, e-commerce, and education.
Whether it is to create different intelligent brains for various avatars, or to provide richer content interaction forms for live broadcasting or e-commerce, Morpheus Cloud can make the production of digital content more automated and efficient. And further bring more interactive and personalized dynamic experience to the development of the virtual world.
Behind this series of products is our understanding and cognition of this matter.
In the past 6 months, we have not only applied cutting-edge technologies such as artificial intelligence to specific scenarios, but also conducted very in-depth research and systematic thinking. At the same time, we have completed research and analysis of nearly 140,000 words, and comprehensively think about games, content, technology and virtual worlds from different angles.
With a strong belief in the trend towards virtual worlds, we recognize that games, while being able to provide open, free, personalised and dynamic digital content, are also the best carriers and experimental sites for access to virtual worlds and Metaverse.
What follows is a comprehensive presentation of the whole process of systematic thinking and a dynamic and evolving perspective on the relationship between people and the world, as well as on the changing trends in the world itself.
We have distilled these 140,000 words of thinking as much as possible, culminating in the long diagram below, which contains the entire framework of our perception of games, Artificial Intelligence and virtual worlds, as well as information on the corresponding key points.
Also, in the descriptions and explanations that follow, we have linked the previously published content to the relevant information in this article by clicking on the jumps for further understanding.
Simply speaking, the ultimate need to be addressed by man as a carbon-based species on Earth is to survive and reproduce. On this basis, we have two solutions: one is to explore outwards, constantly discovering new places to live through interstellar travel, and eventually becoming an interstellar species; the other is to explore inwards, living in virtual worlds in a data-driven way, and eventually becoming a cloud-based species.
In this process, our thinking is structured to be both perceptual and rational, so we use different cognitive systems, including logic, analogy and association, to make sense of the world and ourselves.However, the Gödel incompleteness of mathematics, the quantum uncertainty of physics and the antinomy of philosophy can lead to natural flaws in our logical systems, and thus the inability to understand and cognize the whole system itself. Therefore, if we want to understand the world more fully, we need to obtain information from the outside.
We know that information is the bridge between people and the world, and that its existence is only related to the carrier of the information, the way in which it is combined forms the pathway to our understanding. In the process of disseminating and receiving information, we come into direct contact with it, and also acquire it through interaction with others. Having been exposed to information, we process it using both logic within the cognitive system and analogies and associations outside the system. This brings about not only consensus but also cognitive differences arising from different contexts. It is the existence of such cognitive differences that prevents people from accessing and disseminating information in a consistent and stable manner, with consequences for value creation and development.
As people use information to create value and develop, land provides the space for value creation and technology increases the efficiency of value creation. At the same time, the real world brings obvious limitations from both, with the finite nature of space limiting the area in which people can live and develop, and the finite resources limiting the energy available to people to survive and develop.
Through centuries of exploration and development, information technology has broken the limits of the real world and created virtual worlds based on binary rules. In general, the development of the virtual world and the real world is an intertwined and mutually reinforcing process, and we have divided it into four different parts: the native real world, the actualized virtual world, the digitized real world and the native virtual world.
In the native real world, we use real resources, produce and consume tangible products. The digitized real world emerged as people used digital methods to transfer information from the real world, such as text and images, into the virtual world. At the same time, people began to create and produce directly in the computer, which became the native virtual world. When people took the images and content from the virtual world and presented them with material from the real world, this content and information also formed the actualized virtual world.
In fact, we are still in the early stages of virtual worlds, digitizing the real world while at the same time beginning to create native digital information in virtual worlds. In the future, virtual worlds will evolve into a Metaverse, a virtual world that is always online and at the same time a total collection of information itself, forms of interaction, processes of interaction with each of the digital contents involved. It is a place where countless people can live, socialise, create and experience.
Depending on the development of information technology, we have also divided the Metaverse into different development phases: Metaverse version 0.1: building the basic rules (1940s — 1970s), Metaverse version 0.2: basically transferring information (1980s — 1990s) and Metaverse version 0.3: high frequency information interaction ( 2000s — present). In the future, the Metaverse will also feed back into the real world to an unprecedented degree, and the value that people generate in the real world will be massively invested in the virtual world, where more cycles and iterations will be completed in an economic and social sense.
Currently, games are in fact the best form of access to virtual worlds and Metaverse, and the best carrier for us to build native virtual intelligence. Today’s video games not only allow for high-level graphical interaction, but also provide a degree of logical interaction and a gradual transition from limited options to emergent experiences in terms of plot and narrative.
We also see that more and more games are creating infinite open worlds and emergent interactions with simple rules, but only ABC (AI, Big Data, Cloud) technologies can truly enable dynamic, infinite and personalised interactive experiences in games.
Driven by technology, gaming will usher in an industrial revolution where everything from content production to operations and iteration will happen in the cloud, which will not only increase the efficiency of digital content production, but also bring personalised and dynamic interactions to players and users in real time, thus enabling endless possibilities in the virtual world.
Therefore, the key to creating virtual worlds and metaverse is the automatic generation of intelligent content. We have also divided the production of content into four different stages: specialised content production (PGC), user-produced content (UGC), AI-assisted content generation and fully AI-produced content. The development of technology will provide a rational way of combining information not only on the logical end of the content, but also on the image end, which will provide a rich form of information presentation.
Although the information we receive is, for the most part, visual in the form of images, it is actually the logic system that enables us to perceive the world in a correct and rational way. On this basis, the logic system, together with the intelligent generation of the image content, enables the complete realisation of a dynamic virtual world.
The variety and volume of digital content is so large that it is impractical to tag and implement this logic system by manual means alone, and we need the assistance of artificial intelligence to find the inner logic of content from the vast amount of structured content data. Of all the content domains, games have highly structured data and also have real-time user-content interaction data in different limited scenarios, thus providing a unique advantage for AI to learn and understand the logic of digital content generation. At the same time, games can provide more immersive, free, personalised and dynamically interactive digital content than other forms of content.
This rich content will manifest itself in very unique interactive experiences across media, gaming and social domains. From a media perspective, digital content is inherently medium-based, while technologies such as artificial intelligence will drive more interactive content generation. Current media content is more likely to be generated intelligently outside of the content, for example, recommendation algorithms will constantly recommend the next media content based on the user’s tags. However, we can also see this as a tricky way of interacting. The so-called generation within the content is actually the generation of plots and narratives. In the past, there was no technology that enabled intelligent generation of content internally, but now we can do so with Chaos Box.
It is this dynamic interaction that is generated within the content that gives traditional digital content the interactive properties of a game. At the same time, digital content that can be intelligently generated in response to user interaction, such as intelligent NPCs, dynamic storylines and avatars, will also provide people with new scenarios of corresponding interactions, thus forming a new medium for social relationships in the virtual world.
As a result of this, our society and our lives will enter the next era, an era with more dynamic digital interactive content.
In summary, these are our thoughts on people, thinking, cognition, information, real world, virtual world, Metaverse, games, media, social, content, logic, images, technology, AI, intelligent generation and many other aspects, and they are also our systematic knowledge of what we are doing now.
In fact, whether it is the AI algorithms, cloud service products, visuals we design and build, or the in-depth research analysis and commercialisation partnerships we perform, we are always pursuing true innovation from our bones and constantly iterating on ourselves in repeated thinking.
We believe that with the assistance of Artificial Intelligence, the virtual world will evolve faster and faster. At the same time, digital content will bring a more dynamic, free and open interactive experience to users in the next generation of gaming, media and social.