< Back to The Bohemai Project

Research Report: Investigating the potential for Wave Function Collapse algorithms to model and generate historically accurate, procedurally-generated 3D environments of 1970s San Francisco, using oral histories like Francine Prose's interview as ground truth data.

Executive Summary

This report investigates the feasibility of using Wave Function Collapse (WFC) algorithms to generate historically accurate, procedurally-generated 3D environments representing 1970s San Francisco. The core challenge lies in translating qualitative data, such as oral histories (exemplified by Francine Prose's interview – details assumed for this report), into quantitative data suitable for WFC input. While WFC excels at generating visually coherent and varied environments based on defined constraints, achieving historical accuracy requires a rigorous methodology for data representation and constraint definition. This report explores the potential of this approach, highlighting key developments in WFC, emerging trends in procedural content generation, and potential mitigation strategies to address the challenges inherent in this ambitious undertaking. The accuracy of the generated environment hinges heavily on the quality and quantity of the ground truth data, and the efficacy of its translation into a format compatible with WFC.

Key Developments

Wave Function Collapse has seen significant advancements, moving beyond simple tile-based generation to incorporate more complex features and constraints. Recent developments include:

Emerging Trends

Several emerging trends are relevant to this project:

Mitigation Strategies

Several strategies can mitigate the challenges of using WFC for this project:

Conclusion

Generating a historically accurate 3D model of 1970s San Francisco using WFC and oral histories presents a significant technical challenge, but also a potentially rewarding endeavor. The success of this project depends heavily on the effective structuring and translation of qualitative oral history data into quantitative constraints suitable for WFC. The iterative refinement process and integration of additional procedural generation techniques will be key to achieving both visual fidelity and historical accuracy. Further research into data structuring methods and advanced constraint handling within WFC is crucial for the project's viability.

Sources