FAS Student Papers

Recent advancements in multimodal Generative AI may democratize specialized architectural tasks like interpreting technical drawings and creating 3D CAD models which traditionally require expert knowledge. This paper presents a comparative evaluation study of two systems—GPT-4o and Claude 3.5—in the task of architectural 3D synthesis. It takes as a case study two buildings in Palladio’s Four Books of Architecture (1965): Villa Rotonda and Palazzo Porto. High-level architectural models and drawings of the buildings were prepared inspired by Palladio’s original text and drawing corpus. Through sequential text and image prompting, the study characterizes intrinsic abilities of the systems in (1) interpreting 2D/3D representations of buildings from drawings, (2) encoding the buildings into a CAD software script, and (3) self-improving based on outputs. While both systems successfully generate individual parts, they struggle to accurately assemble these parts into the desired spatial relationships, with Claude 3.5 showing overall better performance, especially in self-correcting its output. The study contributes to ongoing research on benchmarking the strengths and weaknesses of off-the-shelf AI systems in intelligent human tasks requiring discipline-specific knowledge. The results show the potential of language-enabled AI systems to act as collaborative technical assistants in the architectural design process.

Heterochronic parabiosis, a surgical process connecting the circulatory systems of young and old mice, is used to study the regulation of tissue aging and regeneration via protein factors found in blood. Several factors have been shown to improve central nervous system function and ameliorate the effects of aging in mice by increasing stem cell regeneration, but it is less well known which genes are differentially expressed in old mice to induce this “aging reversal” phenotype. We performed an in silico analysis of single cell RNAsequencing (scRNA-seq) data from parabiosed mice to identify transcription factors (TFs) whose expression is significantly dysregulated between old mice and old-young parabiosed mice, which represent the reversed aging state. We employed single-cell regulatory network inference and clustering (SCENIC), a computational framework which identified gene regulatory networks (GRNs) from scRNA-seq data. Regulatory networks from old, old-old parabiosed (as a control), and old-young parabiosed mice were identified from scRNA-seq gene expression matrices spanning 31 cell types. Network activity was then scored by SCENIC as a function of gene expression levels in the networks. These scores formed the basis of comparison of transcription factors between old mice and parabiosed mice. The TFs identified thus far were verified against a set of TFs thought to be dysregulated with parabiosis via differential gene expression analysis. Preliminary results indicated that GRN analysis provides a robust platform for identifying TFs whose expression is significantly altered with parabiosis, and completion of this study will yield a comprehensive understanding of these TFs. In the future, these TFs can serve as therapeutic targets, as compounds can be administered to modulate the expression of these TFs and the genes they regulate in old mice to match the expression profiles identified in old-young parabiosed mice, potentially achieving the aging reversal phenotype.