onformative trains AI to sculpt 3D models from a cube of voxels

What happens when ai becomes the creator?

In the age of technology, inspired by the rising tide of AI powered design and its controversial impact on human creators, onformative’s ‘AI Sculpting’ questions: what happens when AI becomes the creator? How can human creativity and machine-made production feed off each other?

With AI increasingly adapted to design fields, our role as a human creator is changing rapidly. Much faster at learning and reproducing visual output than humans, machines are driven by efficiency, while human creators are primarily motivated by curiosity. Exploring this juxtaposition and curious to find a balance between two co-creators, man and machine, the Berlin-based design studio began an AI research project to program, observe and reinterpret the evolution of a machine learning to sculpt a 3D model. “The central aspect of this co-creation process is that, to a certain degree, we let go of control.” points out the team.

His experiment, presented as a curated series of digital artwork In a climactic 3D generated exhibition environment, it traces the journey of a simple cube transforming into an increasingly recognizable form with each iteration, eventually taking the form of a human form.

AI learns to sculpt 3D models from a cube of voxels, by informational
all images courtesy of information

onformative trains Ai to transform a voxel cube into a statue

Throughout history, man has had a great interest in the manipulation of objects and materials of the moment for aesthetic purposes. Transferring this curiosity to the age of technology and artificial intelligence, design studio onformative designed a machine learning process and AI model as a flagship traditional sculpting tool. “Better experiencing AI as a creator and learning from it became our main pursuit.” points out the team. The AI ​​developed different strategies seeking constant improvement on its way to materializing a certain way. By feeding it different tools, rules, and rewards through reinforcement learning, the team was able to direct the process, but not the outcome, revealing unpredictable end forms.

Set to achieve the given goal of sculpting a 3D model, the AI ​​was trained through reinforcement learning based on rewards and punishments. The agent, defined as a certain machine learning model, was programmed to seek the maximum reward in a process of trial and error. In the voxel-based environment, which started out as a big cube, infinite data and a clear reward structure were provided for the agent to move around. From this initial structure, the agent needed to remove mass to get closer to a predefined target state.

AI learns to sculpt 3D models from a cube of voxels, by informational

With each step, the agent can decide where to go, whether to remove a mass of voxels around it, and how to do it. To make learning possible, it was conditioned in a specific way: when foreign mass was removed, it was rewarded and when mass that should be part of the final sculpture was removed, it was punished. Through trial and error, the agent created his own strategies to achieve the desired shape.

A very surprising result of ‘AI Sculpting’ was the enormous parallelism in the training process, in which the AI ​​was able to simultaneously perform many training sets, resulting in a wide variety of sculptural results. By observing the evolution of the machine’s learning curve, onformative noted the strategies, behaviors, and visual results, continuously experimenting with different parameters and predefined rules.

AI learns to sculpt 3D models from a cube of voxels, by informational

Through the eyes of the machine: AI visualization

Exploring the concept of co-creation further, onformative decided to adopt the agent view to see the process from the creator’s perspective. In an attempt to interpret agent decision making, the design team experimented with displaying AI data such as trust or penalty and reward for individual steps within the 3D environment. By highlighting the agent’s path through the block, for example, they gained another perspective on the creation process itself.

Finally, the team experimented with their creative expression by implementing different tools for the agent to choose from. Through the reinforcement learning setup, this process allowed for greater complexity and resulted in a variety and depth of strategies, as well as visual results. By observing the traces of different tools on the surface of the emerging forms, the sculpting process becomes visible. Various tools have their own unique fingerprint, from rough to fine and fast to slow – each decision the agent made in choosing a different tool or way of targeting had a different result and visual impact. AI learns to sculpt 3D models from a cube of voxels, by informational AI learns to sculpt 3D models from a cube of voxels, by informational

artificial intelligence learns to sculpt 3D models from a cube of voxels, by informational

AI learns to sculpt 3D models from a cube of voxels, by informational

Leave a Reply

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