A breakthrough by a research team from Purdue University's Department of Computer Science and Institute for Digital Forestry, in collaboration with Sören Pirk from Kiel University in Germany, reveals the capability of artificial intelligence (AI) to simulate the intricate growth and shaping of trees in response to their environments.
Inspired by the concept of DNA encoding both tree shape and environmental responses, Professor Bedrich Benes and his team developed innovative AI models that compress the information required for encoding tree forms into a compact, megabyte-sized neural model. The AI models, once trained, encode the local development of trees, enabling the generation of detailed, gigabyte-sized tree models as output.
Published in ACM Transactions on Graphics and IEEE Transactions on Visualizations and Computer Graphics, the research describes the utilization of deep learning, a subset of machine learning within AI, to create growth models for various tree species, including maple, oak, pine, and walnut. The AI models demonstrate the ability to simulate complex behaviors learned from extensive datasets, offering a significant leap in digital tree modeling.
"The AI models learn from large datasets to mimic the intrinsic discovered behavior," says Benes, emphasizing the potential applications in architecture, urban planning, gaming, and entertainment industries where realistic digital tree models are crucial.
The unexpected size reduction of the AI models highlights the efficiency of compressing complex tree behaviors into a minimal amount of data. The researchers anticipate the study to contribute to the development of lab-on-chip devices, combining the chemical fingerprinting capability of this novel SAW-driven biosensor with other acoustic functionalities such as SAW-based mass sensing and microfluidic circuit operations like droplet streaming and mixing.
While the AI models lack training data describing real-world 3D tree geometry, the researchers envision a future where AI can reconstruct 3D geometry data from real trees through digital forestry, aligning with the mission to seamlessly integrate the digital and natural worlds.
More: https://techxplore.com/news/2024-01-ai-simulate-trees-response-environments.html
