HOUND:科学游戏的更多可能 HOUND: More Possibilities of Scientific Game

Introduction

As the game itself expands, game design is unprecedentedly linked with all other areas. Among them, the application of games to scientific research is a typical example. However, looking back at some of these previous games, many rely on the simple labor of a large number of players, such as MIT’s EyeWire, which uses 3D decryption to draw neural network maps.

In the indie game HOUND, players are given more powerful abilities: creating life and their living space. The ultimate goal of the game is to apply it to the study of artificial intelligence, neural networks and ecosystems.

What is HOUND?

Indeed, the original description of the HOUND game is easily linked with Spore. But in fact, the two have a world of difference before.

The developer of HOUND Nick Shesterin studied computer and artificial intelligence, and spent two years practicing the project. He did not design this work as a traditional game, but a powerful simulation engine, equipped with advanced AI system, to help players create their own life forms and ecosystems, accelerate the process of natural selection and genetic mutation to relative Short game time. Through the evolution process of the various forms of life created by the player, the evolution of the organism and the transmission of genes in the natural state are studied.

In the game and on the project website, the author has mentioned such a passage:

Animals, standing behind the door
Getting shot
Falling killed
There was someone, who felt guilty for it
They opened the doors and let the beasts in
The animals entered
And killed everything

Perhaps just like the name of the game “hound” (hound, hunting), the game also implies a deeper reflection from an artificial intelligence researcher.

As a game

Obviously, as a game, the advantages and disadvantages of HOUND are equally obvious.

Extremely distinctive stylized expression

At the beginning of the game, the player is faced with a completely blank space, you need to create your own world from scratch including environmental maps and creatures. At the same time, some presets are provided for import in the game. The most distinctive feature of the game is that it abstracts the mountains, lakes and seas, birds, animals and fish in real life into a collection of chemical-like particles (points) and line segments. As Rowan Crawford of Killscreen magazine puts it, this is an extremely Karl Sims-like visual expression.

Take the example of creating a new creature: When you create a new creature, the default is to create a football-like shape, composed of many particles and line segments. Each particle represents a functional organ trough, and a large number of segments form both the “skeletal morphology” of the organism and the association between the individual particles. Players can adjust the size and shape of the living body through the particle node. The game currently offers a total of 13 different functional particles (organs) for player selection. Players can use these organs instead of the default particle nodes, so that the living body has the corresponding functions, such as energy acquisition, predation, sensory information, and exercise. and many more. At the same time, many organs need to be used in conjunction with other organ particles.

Another important system in the game is the nervous system. After having all kinds of organs, how to make this creature have a normal behavior pattern also requires the player’s thinking and design. The game offered four different neurons, which are different in the neural interface through the logical arrangement of neurons. The organs are linked to control the input and output of the information, thereby determining the behavior pattern of the organism. From the simplest jellyfish creatures to life with more complex behavior patterns. With the powerful editing capabilities provided by the game, players can implement it in their own world. Predators and prey, even alien invasive species, may appear in the player’s ecosystem.

At the same time, HOUND also provides a worldwide network of gene transmissions, which means that the genetic information of different players will spread with the Internet, resulting in a larger and more complex system. As part of a research project, researchers can conduct further research by collecting and researching the data generated during the player’s game.

Design flaw

It is not difficult to see from the above description that as a game, HOUND is really not a “game”. First of all, the relatively difficult scientific background of the game is enough to be prohibitive. In the process of understanding the game, a lot of biological terms and theories and the more cumbersome and meticulous operation process are really dizzying (Considering the game is in English only, the problem is obviously more serious for player whose native language is not English).

Secondly, as a simulation game, if you don’t have a deep understanding of the game, it will easily fall into a situation where you are overwhelmed. The author himself said that HOUND has no concept of achievement or clearance. The only goal is the survival of an organism or the operation of the ecosystem. And if you are not happy, you can also destroy it directly. And this obviously requires a lot of patience.

Last but not least, HOUND’s interaction model is really unfriendly. On the basis of the abstraction of the game screen itself, a large number of operations of the game are collected into the left and right dashboards. In a small area, the mouse has to complete various operations such as selection, stretching, and scrolling. The entire UI interface of the game also lacks recognition. For this reason, the developer has also set a button to let the player check the help at any time. From the beginning of the tutorial, the steep learning curve has greatly eroded the player’s desire to play, not to mention the deep experience and pleasure.

Future prospects

In my communication with developers, he also mentioned expectations about the progress of future projects. In response to the problem of getting started and understanding the game, the developer plans to further decentralize the teaching part of the game, giving players a greater choice of learning pace. The flow of genes will become more visual and dynamic, and system AI will truly combine and reconstruct the genes created by players to produce entirely new species and have a greater impact on the global ecosystem. At the same time, players may be given more authority and freedom, including observing the flow of genes they design in the ecosystem and the more versatile interaction between players. Despite the design problems, there is no doubt that HOUND shows us the wider possibilities of games.

引言

随着游戏本身外延的扩张,游戏设计正在和其他各个领域前所未有地联系在一起。其中,将游戏运用于科学研究就是一个典型的例子。不过回顾先前的一些此类游戏,很多还只是依靠大量玩家的简单劳动,比如 MIT 的 EyeWire,使用3d解密的形式来绘制神经网络图。
而像 CMU 和 Stanford 所开发的的 EteRNA 和它的原型 Foldit 则更进一步,让玩家来自己设计RNA的构造,帮助科研人员研究 RNA 的展开。

更多的科学游戏可以从这里找到。

在独立游戏 HOUND 中,玩家们则被给予了更为强大的能力:创造生命,以及它们的生存空间。而这款游戏的最终目的,则是将其运用于人工智能,神经网络和生态系统的研究中。

什么是 HOUND?

的确,对于 HOUND 这款游戏最初的描述,很容易让人联想起 Spore(孢子)。但实际上这两者之前有着天壤之别。

HOUND 的开发者 Nick Shesterin 大学所学的专业是计算机与人工智能,他花费了两年的时间来实践这个项目。他没有把这个作品作为传统的游戏来设计,而是一个强大的模拟引擎,搭载着先进的 AI 系统,帮助玩家创作出自己的生命形态和生态系统,将自然选择和基因突变的过程加速到相对短暂的游戏时间中。通过玩家所创作的千姿百态的生命形式的演化过程,来研究自然状态下生物的演变和基因的传递。

在游戏中和项目网站,作者都提到了这样一段话:

Animals, standing behind the door
​Getting shot
Falling killed
​There was someone, who felt guilty for it
​They opened the doors and let the beasts in
The animals entered
​And killed everything

或许正如同游戏的名字“hound”(猎犬、追猎),游戏也隐含了作为人工智能研究者更深层次的思考。

作为游戏的 HOUND

显然,作为一款游戏来说,HOUND 的优点和缺点同样明显

极其鲜明的风格化表达

在游戏的一开始,玩家所面对的是一个完全空白的空间,你需要从零开始创造自己的世界。包括环境地图和生物。同时,游戏中也提供了一些预设以供导入。而游戏最有特色的地方就在于,它将现实生活中的山河湖海,鸟兽鱼虫,全部抽象成了类似化学结构式的颗粒(点)和线段的集合。正如 Killscreen 杂志的 Rowan Crawford 所说,这是一种极其 Karl Sims 式的视觉表达。以创造一个新生物为例:当你新建一个新生物的时候,默认创建的是一个类似足球的形体,由很多颗粒和线段组成。每一个颗粒,都代表着一个功能器官槽位,而大量的线段既构成了生物的“骨骼形态”同时也表现出各个颗粒之间的关联。玩家可以通过颗粒节点来调整生命体的大小形态。游戏目前一共提供了 13 中不同的功能颗粒(器官)来共玩家选择,玩家可以用这些器官来代替默认的颗粒节点,从而让生命体具备相应的功能,比如获取能量、捕食、感知信息、运动等等。同时,很多器官需要与其他器官颗粒配合使用。游戏中另外一个重要的系统就是神经系统。在拥有了各种器官之后,如何让这个生物有一个正常的行为模式同样需要玩家的思考与设计游戏中一共有四种不同的神经元,在神经界面中通过神经元的逻辑排布,将不同的器官联系起来,控制信息的输入与输出,从而决定该种生物的行为模式。从最简单的水母式的生物,到拥有更复杂行为模式的生命。通过游戏提供的强大编辑能力,玩家都可以在自己的世界中将其实现。捕食者和被捕食者,甚至外来入侵物种,都可能出现在玩家的生态系统之中。

与此同时,HOUND 还提供了一个世界范围内的基因传播网络,这也就意味着不同玩家的基因信息会随着互联网进行扩散,从而形成一个更广阔和复杂的大型系统。作为一项研究计划的一部分,通过收集和研究玩家游戏过程中所产生的数据,研究人员可以对其进行进一步研究。

设计的缺陷

从上面的描述不难看出,作为一款游戏,HOUND 实在是太不“游戏”了。首先,游戏相对艰深的科学背景就足以让人望而却步,在理解游戏的过程中,大量的生物学术语和理论以及较为繁琐细致的操作流程确实令人头晕(在只有英文的情况下对于母语非英语的玩家这一问题显然更为严重)。

其次,作为一款模拟养成类游戏,如果对于游戏没比较深入的了解,很容易就会陷入不知所措的境地中。作者自己也说,HOUND 并没有成就或者过关的概念,唯一的目标就是一个生物体的生存与否,生态系统的运转情况。而如果你不开心,也可以直接将其毁于一旦。而这显然需要大量的耐心。

最后但尤为突出的一点就是,HOUND 的交互模式实在是不友好。在本身游戏画面就很抽象的基础之上,游戏大量的操作被集合到了左右两个仪表盘中,在很小的区域内鼠标要完成选中,拉伸、滚轮等多种操作。游戏整个的UI界面也缺乏识别度,为此,开发者还专门设置了一个按键,来让玩家随时查阅帮助。从教程开始,陡峭的学习曲线就极大地消磨了玩家的游玩欲望,更不用说深层次的体验和愉悦感。

未来的展望

在和开发者的交流中,他也提到了关于未来项目进展的预期。针对游戏上手和理解难度较大的问题,开发者计划对游戏的教学部分进行进一步的分散化,让玩家有更大的选择学习节奏的余地。基因的流动会变得更加可视化和动态化,系统 AI 将会真正将玩家所创造的基因组合和重构,产生出全新的物种,并对全球生态系统产生更加重大的影响。同时,玩家可能会被赋予更大的权限和自由度,包括观察自己设计的基因在生态系统中的流动以及更急加多样化的玩家间互动等。尽管设计存在问题,但无疑,HOUND 为我们展示了游戏更广泛的可能性。

Leave a Reply

Your email address will not be published. Required fields are marked *