The field of generative design and evaluation is rapidly evolving, with a focus on developing innovative methods for sustainable building design and visual composition. Recent research has explored the use of energy-aware and function-feasible generative frameworks, such as GreenPlanner, to accelerate design evaluation and generation. Other studies have proposed frameworks for large-scale thermal building data generation, such as BuilDa, to drive machine learning research. Noteworthy papers that have advanced the field include GreenPlanner, which achieves over 10^5 times acceleration in evaluation with high accuracy, and PaCo-RL, which introduces a comprehensive framework for consistent image generation with pairwise reward modeling. Additionally, SA-IQA redefines image quality assessment for spatial aesthetics with multi-dimensional rewards, setting a new standard for spatial aesthetics evaluation.
Advancements in Generative Design and Evaluation for Sustainable Buildings and Visual Composition
Sources
GreenPlanner: Practical Floorplan Layout Generation via an Energy-Aware and Function-Feasible Generative Framework
A Highly Configurable Framework for Large-Scale Thermal Building Data Generation to drive Machine Learning Research
PSR: Scaling Multi-Subject Personalized Image Generation with Pairwise Subject-Consistency Rewards
Generative Adversarial Gumbel MCTS for Abstract Visual Composition Generation
EvalTalker: Learning to Evaluate Real-Portrait-Driven Multi-Subject Talking Humans
Prescriptive tool for zero-emissions building fenestration design using hybrid metaheuristic algorithms
PaCo-RL: Advancing Reinforcement Learning for Consistent Image Generation with Pairwise Reward Modeling
SA-IQA: Redefining Image Quality Assessment for Spatial Aesthetics with Multi-Dimensional Rewards