for latest works and updates please visit 
tugrulsalci ︎
distraktion lab ︎

   

Cybetnetic Flora by Distraktion

2023
Realtime Generative AI Installation
Microcosmic Biosphere
The artwork Cybernetic Flora exemplifies the intricate merging of elements drawn from diverse botanical sources, creating a symbiosis of forms that seem to transcend their origins. Inspired by the complex interrelationships found in natural ecosystems, the piece combines features reminiscent of marine and terrestrial life, such as jellyfish, coral, amber, and dandelion, into a cohesive and harmonious entity.
Through the lens of Deleuze and Parnet, this fusion is not merely an imitation but a dynamic assemblage that reshapes our understanding of existence. Their insights into the fabric of interconnectivity resonate deeply in the creation of Cybernetic Flora, where the realms of artificial intelligence, cyber-ethics, and human-AI co-evolution interweave, forming a unique plane of immanence. This realm is a manifestation of Deleuze and Parnet's vision, where the distinction between the 'natural' and 'artificial' dissolves, unveiling a domain constructed through the dynamic interplay of varied forces and entities.

Further enriching this assemblage, Michel Serres' insights into the transformative role of 'parasitic' noise within these interactions add a profound layer of complexity, transforming disruptions into catalysts for novel forms of existence and co-functioning. In this context, the generative space of Cybernetic Flora emerges as a manifestation of these theoretical perspectives, where disparate organic forms and digital creativity engage in a non-linear dialogue. The artwork thus becomes a dynamic participant in the broader conversation about life's interconnectedness, embodying the principle of symbiosis in both form and function.



Methodology
Lindenmayer's foundational work highlighted the system's utility in theoretical biology, where it serves to model inputs received by cells from both directions along a filament, dictating changes in state and output based on the present state and two inputs it receives, which is crucial for branching filaments as well. By encoding biological processes into a formalized mathematical framework, researchers have been able to simulate and analyze a wide array of growth phenomena, from the branching patterns of trees to the spiral arrangements of leaves. This inherent capacity for abstraction and representation of biological structures has extended the utility of the L-system beyond its origins in theoretical biology, finding applications in various fields including computer graphics, generative art, and procedural content generation.






Stable Diffusion, a pivotal model in visual arts, introduces an innovative approach to image denoising and enhancement. By using diffusion models and applying them in the latent space of powerful pretrained autoencoders, Stable Diffusion excels in removing noise while retaining essential structural details. This process not only refines images through an iterative scheme to improve visual fidelity but also incorporates cross-attention layers, enhancing the model's flexibility for various tasks like text-to-image synthesis, unconditional image generation, and super-resolution. In the realm of creative media, fusing L-system with Stable Diffusion—the technique referred to as "Denoising Path Orientation"—utilizes the model's denoising strengths to precisely refine L-system-generated patterns. This fusion not only improves visual quality but also adopts the L-system to use in convolution processes within AI models, thereby broadening its interdisciplinary applications.

© Tuğrul Şalcı 2023