Direct sound printing (DSP) is a new additive manufacturing process which utilizes sound waves as a driving energy source instead of the conventional heat or light-based methods. This is done through inducing sonochemical reactions in localized cavitation regions using highly focused ultrasound. Advantages to these methods include a wider range of materials to print with, greater energy penetration depth allowing for more isotropic prints with fewer layers, and remote distance printing (RDP), where solid material can be cured across opaque barriers or through walls. While most thermoset polymers require post-process curing time, experiments notably using Polydimethylsiloxane (PDMS) resin demonstrated near-instant formation, skipping the process entirely. Holographic Direct Sound Printing (HDSP) also works as a method to improve print speed and efficiency using acoustic holography, elevating the previously voxel-by-voxel one-dimensional process to multiple dimensions while still using a single transducer. Applications for DSP include remote in-body printing (implants) and composite materials featuring metallic particles. Studies of DSP prints using sonochemiluminescence (SCL) showed near identical material properties to thermally cured counterparts, highlighting the expanding capabilities of sound-based printing systems that rival existing technology.

Large Language Models (LLMs) are machine-learning models that are trained to understand, predict, and generate human-like text by learning statistical patterns and contextual relationships within language. In recent years, the rapid advancement of LLMs has significantly impacted several areas within artificial intelligence, particularly in natural language processing (NLP) and computer program analysis. LLM-based computer program analysis involves tasks such as error detection, code generation, and identifying structural similarities between pieces of code. These tasks leverage the models’ ability to interpret programming languages as a form of structured text, enabling new approaches to analyzing and understanding code. This research project aims to explore and evaluate the capability of LLMs in assessing the semantic similarity between pairs of Python functions. Specifically, it seeks to determine how effectively an LLM can recognize when two pieces of code perform the same or similar functions, even if their syntax or structure differs. This kind of semantic comparison is useful in applications such as plagiarism detection, code recommendation systems, and automated feedback in computer programming education. The current approach uses the Abstract Syntax Tree (AST) to extract the features of each function, which is then used as input to an LLM trained on code to get the embeddings. The results from both AST structural similarity and LLM embeddings are then combined to calculate the final semantic similarity between two functions.

Despite our technologically-focused society, research has shown that the rise of digital games has not forced analog games to cease existing. This suggests that there may be factors in analog games that may not be transferable to the digital versions. Literature shows other tools currently exist for solely digital games or analog games, but none that can measure both effectively and accurately. The Games Experience Questionnaire (GEQ) is a commonly used user experience (UX) tool while playing a digital game, but may be adaptable to measure experience playing analog games. This project seeks to answer the following questions: Which factors of flow (simply defined as an in-the zone state) are affected the most when converting a board game into digital format and vice versa? What changes could be made (in terms of elements of flow) to make them equally or more enjoyable? How does this insight allow new UX tools to be developed to measure enjoyment or flow? This mixed-methods study intends to investigate these questions by having participants play analog and digital versions of UNO and Tetris and later complete a modified version of the GEQ. This can help determine whether flow was achieved and understand the positive and negative areas of their experience. The goal of this project is to give insight for game developers into how they could conserve the essence of the original version of a game when translating it into another format to provide consistent or improved enjoyment.

As robotic and automated systems (RAS) become more common in natural resource management, understanding public acceptance of these technologies in nature-based settings is important. This study investigates whether prior use of RAS influences acceptance of low-cost, open-source autonomous underwater vehicle (LoCO AUV) designed to monitor water quality and aquatic invasive species (AIS). A representative survey of Minnesotans assessed how often respondents used select automated technologies (e.g., robotic vacuums, drones), how acceptable they found LoCO AUV’s use in monitoring, and the relationship between the two. Results showed that LoCO acceptance for water quality monitoring varied across levels of prior technology experience: those with low to moderate levels of technology use were more accepting than those with no experience, while those with greater experience were significantly less so. A similar, but non-significant finding was present in the relationship between technology use and acceptance of robotic AIS monitoring. Findings suggest that some exposure to technology may encourage more acceptance of RAS in outdoor settings, but at some point, exposure leads to significantly less acceptance. This study contributes to a growing understanding of the perceived acceptance of RAS in nature-based settings by highlighting the relationship between technology familiarity and public attitudes. Further, it emphasizes the need to consider public experience when integrating RAS into these settings, as it varies and is complex.