BIM-Edit: Benchmarking Large Language Models for IFC-Based Building Information Modeling

Large language models (LLMs) are increasingly applied to computer-aided design (CAD) to generate design artifacts from textual instructions. In engineering practice, this requires more than creating new geometry, models must also understand existing scenes, edit them correctly, and preserve semantics and relations. However, many CAD benchmarks focus on creating new models rather than editing existing ones, and mostly evaluate geometric correctness. We introduce BIM-Edit, a benchmark for evaluating LLMs on natural-language editing of Building Information Models (BIM) represented in the Industry Foundation Classes (IFC) format. BIM provides a challenging testbed because building models encode geometry together with semantic and relational structure. BIM-Edit contains 324 editing tasks spanning 11 realistic building models and 36 synthetic scenes. Tasks are expressed using three instruction categories - direct, spatial, and topological - covering both explicit and scene-grounded edits. We evaluate outputs along three dimensions: geometric accuracy, semantic validity, and topological consistency. Across evaluated LLMs, the best-performing model achieves only 49.5% average score across the three metrics, and no model fully solves more than 3.4% of tasks. These results demonstrate a substantial gap between current LLM capabilities and the requirements of structured engineering design workflows.

Transductively Informed Inductive Program Synthesis

Abstraction and reasoning in program synthesis has seen significant progress through both inductive and transductive paradigms. Inductive approaches generate a program or latent function from input-output examples, which can then be applied to new inputs. Transductive approaches directly predict output values for given inputs, effectively serving as the function themselves. Current approaches combine inductive and transductive models via isolated ensembling, but they do not explicitly model the interaction between both paradigms. In this work, we introduce \acs{tiips}, a novel framework that unifies transductive and inductive strategies by explicitly modeling their interactions through a cooperative mechanism: an inductive model generates programs, while a transductive model constrains, guides, and refines the search to improve synthesis accuracy and generalization. We evaluate \acs{tiips} on two widely studied program synthesis domains: string and list manipulation. Our results show that \acs{tiips} solves more tasks and yields functions that more closely match optimal solutions in syntax and semantics, particularly in out-of-distribution settings, yielding state-of-the-art performance. We believe that explicitly modeling the synergy between inductive and transductive reasoning opens promising avenues for general-purpose program synthesis and broader applications.

Shedding Light in Task Decomposition in Program Synthesis: The Driving Force of the Synthesizer Model

Task decomposition is a fundamental mechanism in program synthesis, enabling complex problems to be broken down into manageable subtasks. ExeDec, a state-of-the-art program synthesis framework, employs this approach by combining a Subgoal Model for decomposition and a Synthesizer Model for program generation to facilitate compositional generalization. In this work, we develop REGISM, an adaptation of ExeDec that removes decomposition guidance and relies solely on iterative execution-driven synthesis. By comparing these two exemplary approaches-ExeDec, which leverages task decomposition, and REGISM, which does not-we investigate the interplay between task decomposition and program generation. Our findings indicate that ExeDec exhibits significant advantages in length generalization and concept composition tasks, likely due to its explicit decomposition strategies. At the same time, REGISM frequently matches or surpasses ExeDec's performance across various scenarios, with its solutions often aligning more closely with ground truth decompositions. These observations highlight the importance of repeated execution-guided synthesis in driving task-solving performance, even within frameworks that incorporate explicit decomposition strategies. Our analysis suggests that task decomposition approaches like ExeDec hold significant potential for advancing program synthesis, though further work is needed to clarify when and why these strategies are most effective.

AbstractBeam: Enhancing Bottom-Up Program Synthesis using Library Learning

LambdaBeam is a state-of-the-art execution-guided algorithm for program synthesis that incorporates higher-order functions, lambda functions, and iterative loops into the Domain-Specific Language (DSL). LambdaBeam generates every program from the start. Yet, many program blocks or subprograms occur frequently in a given domain, e.g., loops to traverse a list. Thus, repeating programs can be used to enhance the synthesis algorithm. However, LambdaBeam fails to leverage this potential. For this purpose, we introduce AbstractBeam: A novel program synthesis framework that employs Library Learning to identify such program repetitions, integrates them into the DSL, and thus utilizes their potential to boost LambdaBeam's synthesis algorithm. Our experimental evaluations demonstrate that AbstractBeam significantly improves LambdaBeam's performance in the LambdaBeam integer list manipulation domain. Additionally, AbstractBeam's program generation is more efficient compared to LambdaBeam's synthesis. Finally, our findings indicate that Library Learning is effective in domains not specifically crafted to highlight its benefits.