Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Generating Physically Sound Designs from Text and a Set of Physical Constraints
Authors: Gregory Barber, Todd Henry, Mulugeta Haile
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate TIDES on a series of structural optimization problems operating under different load and support conditions, at different resolutions, and experimentally in the lab by performing the 3-point bending test on 2D beam designs that are extruded and 3D printed. |
| Researcher Affiliation | Industry | Gregory Barber, Todd C. Henry, Mulugeta A. Haile DEVCOM Army Research Laboratory Aberdeen Proving Ground, MD 21005 EMAIL |
| Pseudocode | No | The paper describes the TIDES framework and its components using textual descriptions and a diagram (Figure 2), but does not include any explicit pseudocode blocks or algorithms. |
| Open Source Code | Yes | Question: Does the paper provide open access to the data and code, with sufficient instructions to faithfully reproduce the main experimental results, as described in supplemental material? Answer: [Yes] Justification: We describe the experimental configurations in Appendix B and we will make our code available. |
| Open Datasets | No | The paper focuses on generating designs and conducting experiments on those generated designs (e.g., 3D printed beams). It does not explicitly state the use of any external, publicly available datasets for its methodology, nor does it provide access information for such datasets. |
| Dataset Splits | No | The paper does not describe using any pre-existing datasets with traditional training, validation, or test splits. The work involves generating and testing designs rather than evaluating on standard partitioned datasets. |
| Hardware Specification | Yes | Question: For each experiment, does the paper provide sufficient information on the computer resources (type of compute workers, memory, time of execution) needed to reproduce the experiments? Answer: [Yes] Justification: The paper does not require extensive computing resources. All runs and trials were performed on a computer with an RTX A4000, 32GB Ram, and an Intel i7 processor. |
| Software Dependencies | No | The paper mentions using 'pytorch', 'Torch Vision library [35]', 'Auto Grad [36] SIMP implementation from [8]', and 'Adam W [37] optimizer'. However, specific version numbers for PyTorch and Torch Vision are not provided, which are necessary for a reproducible description of ancillary software. |
| Experiment Setup | Yes | In Appendix B we provide descriptions of all experimental setups and the hyperparameters used for each run. ... In the loss function β1 = 50 and β2 = 100. The Adam W [37] optimizer was used with a learning rate of 0.25 and train for 100 epochs. |