'HOME'CAST: Shaping the Built Environment through Data-Driven Innovation | Student Presentations

Item: 'HOME'CAST: Shaping the Built Environment through Data-Driven Innovation | Student Presentations
Rating: 4
Author: PostSingularity Institute

by Harvard GSD

Applied Sciences & Engineering Arts & Architecture AMVLandscape architecture and design AMVDCity and town planning: architectural aspects

🎙 Harvard GSD 👥 124K 📅 October 10, 2025 ⏱ 50 min 👁 117 📄 conference presentation 🧭 2026-08-16

Available in: English (current) Français

Keywords

building simulationthermal storagedemand responsehumidity controlreinforcement learning

Summary

This video features three doctoral student presentations from the ‘HOME’CAST conference at Harvard GSD, focusing on data-driven innovation in the built environment. The first presentation by Yiwei (Lucy) Lyu explores harnessing building thermal storage to respond to time-varying emission signals, aiming to reduce carbon emissions through demand-responsive control. She presents preliminary results from simulations on a highly efficient house and city-scale models, highlighting potential energy savings and trade-offs with occupant comfort. The second presentation by Sang Won Kang addresses humidity management in buildings, proposing a data-driven approach using dense sensing, predictive modeling, and adaptive control based on reinforcement learning and Bayesian optimization. He emphasizes the limitations of relative humidity as a metric and the need for absolute metrics and context-aware control. The third presentation by Lyna (likely a typo for ‘Lyna’ but not detailed in the transcript) is not covered in the provided transcript. Overall, the video showcases cutting-edge research at the intersection of building science, AI, and sustainability, with a focus on practical applications and real-world validation.

168 words

Critical Evaluation

Value of the Information & Strength of the Argument

The presentations offer valuable insights into advanced building technologies and data-driven methods. Lucy’s work provides a clear argument for using time-varying emission signals to optimize building operations, supported by preliminary simulation results. Sang’s presentation makes a compelling case for rethinking humidity control, backed by a well-structured research framework and preliminary studies. The argumentation is solid, with each presenter clearly stating their hypotheses, methodologies, and expected outcomes. However, as these are works in progress, the results are preliminary and not yet validated in real-world settings.

Scientific Rigor, Source Quality, Title Accuracy

The presentations demonstrate rigorous scientific methodology, with clear research questions, hypotheses, and planned validation steps. The sources cited are primarily academic and institutional, such as the Harvard Center for Green Buildings and Cities and the Comstock residential building models. The title accurately reflects the content, focusing on data-driven innovation in the built environment. The video includes a brief introduction and Q&A session, but no commercial advertisements were noted. The overall quality is high, though the lack of peer review and the preliminary nature of the results temper the reliability.

188 words

Title / Content Match

The title accurately reflects the content, which focuses on data-driven innovation in the built environment through student research presentations.

Quality & Reliability

7/10

Presentations are from doctoral candidates at a reputable institution, with clear methodology and preliminary results. However, they are works in progress and not peer-reviewed.

Chapters

Cited Sources

Harvard Center for Green Buildings and Cities — Mentioned as the lab where Lucy's research is based (House Zero).
Comstock residential building models — Used for city-scale simulations in Lucy's research.

Concurring Sources

Harvard Center for Green Buildings and Cities — Supports the research context for building sustainability.

Contribution & Novelties

This video provides a unique glimpse into ongoing doctoral research at Harvard GSD, showcasing innovative approaches to building design and operation. Lucy’s work on using thermal storage for demand response is novel in its integration of real-time emission signals and city-scale analysis. Sang’s research on humidity control with AI and adaptive algorithms addresses a often-overlooked aspect of building performance. The presentations highlight the importance of bridging simulation and real-world validation, and they offer new perspectives on designing for a low-carbon future.

Pour aller plus loin :

Demand response — Relevant to Lucy’s work on load shifting.
Model predictive control — Central to Sang’s control framework.
Reinforcement learning — Used in Sang’s adaptive control.
Thermal mass — Key concept in Lucy’s research on thermal storage.

124 words

Radar Profile

The radar profile shows high scores in technical level and information quality, reflecting the advanced nature of the research. The lower scores in quantity and reliability are due to the preliminary stage of the work and the limited number of sources cited.

Reliability 7/10