We present AircraftVerse, a publicly available aerial vehicle design dataset. Aircraft design encompasses different physics domains and, hence, multiple modalities of representation. The evaluation of these designs requires the use of scientific analytical and simulation models ranging from computer-aided design tools for structural and manufacturing analysis, computational fluid dynamics tools for drag and lift computation, battery models for energy estimation, and simulation models for flight control and dynamics. AircraftVerse contains $27{,}714$ diverse air vehicle designs - the largest corpus of designs with this level of complexity. Each design comprises the following artifacts: a symbolic design tree describing topology, propulsion subsystem, battery subsystem, and other design details; a STandard for the Exchange of Product (STEP) model data; a 3D CAD design using a stereolithography (STL) file format; a 3D point cloud for the shape of the design; and evaluation results from high fidelity state-of-the-art physics models that characterize performance metrics such as maximum flight distance and hover-time. We also present baseline surrogate models that use different modalities of design representation to predict design performance metrics, which we provide as part of our dataset release. Finally, we discuss the potential impact of this dataset on the use of learning in aircraft design, and more generally, in the emerging field of deep learning for scientific design. AircraftVerse is accompanied by a datasheet as suggested in the recent literature, and it is released under Creative Commons Attribution-ShareAlike (CC BY-SA) license. The dataset with baseline models are hosted at http://doi.org/10.5281/zenodo.6525446, code at https://github.com/SRI-CSL/AircraftVerse, and the dataset description at https://uavdesignverse.onrender.com/.

Advances in artificial intelligence create new opportunities for computers to support humans as peers in hybrid teams in several complex problem-solving situations. This paper proposes a decision-making architecture for adaptively informing decisions in human-computer collaboration for large-scale competitive problems under dynamic environments. The proposed architecture integrates methods from sequence learning, model predictive control, and game theory. Computers in this architecture learn objectives and strategies from experimental data to support humans with strategic decisions while operational decisions are made by humans. The paper also presents data-driven methods for partitioning tasks among a team of computers in this architecture. The generalized methodology is illustrated on the real-time strategy game Starcraft II. The results from this application show that low-performing players can benefit from the game-theoretic decision support whereas this support can be overly conservative for high-performing players. The proposed approach provides safe though suboptimal suggestions particularly against an opponent with an unknown level of expertise. The results further show that problem solution with a team of computers based on non-intuitive task partitioning significantly improves the quality of decisions compared to an all-in-one solution with a single computer.

This research paper presents recent work exploring the power of natural language processing (NLP) methods applied to qualitative engineering education data. As NLP and other machine learning methods are developed for qualitative data, it is important to prioritize the role that theory plays in rigorous qualitative research, where the selection of a theoretical framework serves as the lens by which the research project is framed, results are analyzed, and findings are brought to light. Indeed, the view from a different theoretical lens can highlight novel or new findings. In this work, we seek to explore the viability of NLP methods for helping researchers select appropriate frameworks. In this work, we present our method to train a Python-based NLP algorithm to analyze an existing data set of interview data using one theoretical lens: Community of Practice theory, an oft-used theory in graduate education literature, which is the topic of the interview corpus to investigate. We present and test two methods for developing dictionaries by which to train the algorithm: An expert-curated dictionary and a machine-generated dictionary compiled by mining the theoretical framework sections of published literature employing Community of Practice theory. We apply these two dictionaries to analyze a corpus of 54 interview transcripts investigating graduate engineering attrition. The high dimensional data from NLP can be compared using Principal Component Analysis (PCA) visualization and pairwise distance plots to determine which method results in the most well-defined structure indicating agreement between the dictionary and the corpus of interview transcripts. In the discussion, we highlight opportunities for using these automated methods to help researchers with qualitative data analysis and warn against potential dangers and ethical ramifications for using machine learning and NLP for social science data. This work will have impact on the disciplinary communities working to embed computational language-based methods into engineering education research, and for the qualitative methods communities across social science and education disciplines.

This research full paper proposes opportunities to expand qualitative and textual data analysis using natural language processing (NLP), and demonstrates opportunities to use NLP in engineering education work in our presentation of two NLP-based projects as examples of how NLP can be used. The discipline of engineering education frequently employs qualitative data analysis techniques, but fundamentally, the analysis of large corpora of textual documents is limited by researcher's time. In this paper, we present a brief review literature of how NLP has been used in other qualitative research fields. This portion of the paper is aimed to provide a clear description of NLP for those unfamiliar with machine learning and natural language processing methods. The second part of the paper will provide two brief examples of how NLP is being employed in our research group. Example 1 is a study of engineering résumés, with the intention of being able to calculate the “disciplinary discourse density” based off the engineering language presented in engineering résumés, a technique validated in prior qualitative studies by hand. Example 2 is a genre analysis of engineering literature reviews, seeking to understand the ways in which sentences linguistically build into arguments, such that the task of writing literature reviews might be demystified. This paper will have a methodological impact for researchers attempting to use NLP methods to analyze qualitative data, articulating the opportunities and barriers in using these methods for engineering education research.