Barrett, The Honors College Thesis/Creative Project Collection

The research presented in this Honors Thesis provides development in machine learning models which predict future states of a system with unknown dynamics, based on observations of the system. Two case studies are presented for (1) a non-conservative pendulum and (2) a differential game dictating a two-car uncontrolled intersection scenario. In the paper we investigate how learning architectures can be manipulated for problem specific geometry. The result of this research provides that these problem specific models are valuable for accurate learning and predicting the dynamics of physics systems.<br/><br/>In order to properly model the physics of a real pendulum, modifications were made to a prior architecture which was sufficient in modeling an ideal pendulum. The necessary modifications to the previous network [13] were problem specific and not transferrable to all other non-conservative physics scenarios. The modified architecture successfully models real pendulum dynamics. This case study provides a basis for future research in augmenting the symplectic gradient of a Hamiltonian energy function to provide a generalized, non-conservative physics model.<br/><br/>A problem specific architecture was also utilized to create an accurate model for the two-car intersection case. The Costate Network proved to be an improvement from the previously used Value Network [17]. Note that this comparison is applied lightly due to slight implementation differences. The development of the Costate Network provides a basis for using characteristics to decompose functions and create a simplified learning problem.<br/><br/>This paper is successful in creating new opportunities to develop physics models, in which the sample cases should be used as a guide for modeling other real and pseudo physics. Although the focused models in this paper are not generalizable, it is important to note that these cases provide direction for future research.

High-entropy alloys possessing mechanical, chemical, and electrical properties that far exceed those of conventional alloys have the potential to make a significant impact on many areas of engineering. Identifying element combinations and configurations to form these alloys, however, is a difficult, time-consuming, computationally intensive task. Machine learning has revolutionized many different fields due to its ability to generalize well to different problems and produce computationally efficient, accurate predictions regarding the system of interest. In this thesis, we demonstrate the effectiveness of machine learning models applied to toy cases representative of simplified physics that are relevant to high-entropy alloy simulation. We show these models are effective at learning nonlinear dynamics for single and multi-particle cases and that more work is needed to accurately represent complex cases in which the system dynamics are chaotic. This thesis serves as a demonstration of the potential benefits of machine learning applied to high-entropy alloy simulations to generate fast, accurate predictions of nonlinear dynamics.

Robots are often used in long-duration scenarios, such as on the surface of Mars,where they may need to adapt to environmental changes. Typically, robots have been built specifically for single tasks, such as moving boxes in a warehouse or surveying construction sites. However, there is a modern trend away from human hand-engineering and toward robot learning. To this end, the ideal robot is not engineered,but automatically designed for a specific task. This thesis focuses on robots which learn path-planning algorithms for specific environments. Learning is accomplished via genetic programming. Path-planners are represented as Python code, which is optimized via Pareto evolution. These planners are encouraged to explore curiously and efficiently. This research asks the questions: “How can robots exhibit life-long learning where they adapt to changing environments in a robust way?”, and “How can robots learn to be curious?”.

Is there a mismatch between urban farmers’ perceptions of their farm’s environmental sustainability and its actual environmental impact? Focusing on the use of water and nutrients on each farm as described by the farmers through interviews, it is evident that there is some level of disconnect between ideals and practices. This project may aid in bridging the gap between the two in regard to the farmers’ sustainability goals. This project will move forward by continuing interviews with farmers as well as collecting soil and water from the farms in order to more accurately quantify the sustainability of the farms’ practices. This project demonstrates that there is some degree of misalignment between perception and reality. Two farms claimed they were sustainable when their practices did not reflect that, while 2 farms said they were not sure if they were sustainable when their practices indicated otherwise. Samples from two farms showed high concentrations of nutrients and salts, supporting the idea that there may be a mismatch between perceived and actual sustainability.

Colorimetric assays are an important tool in point-of-care testing that offers several advantages to traditional testing methods such as rapid response times and inexpensive costs. A factor that currently limits the portability and accessibility of these assays are methods that can objectively determine the results of these assays. Current solutions consist of creating a test reader that standardizes the conditions the strip is under before being measured in some way. However, this increases the cost and decreases the portability of these assays. The focus of this study is to create a machine learning algorithm that can objectively determine results of colorimetric assays under varying conditions. To ensure the flexibility of a model to several types of colorimetric assays, three models were trained on the same convolutional neural network with different datasets. The images these models are trained on consist of positive and negative images of ETG, fentanyl, and HPV Antibodies test strips taken under different lighting and background conditions. A fourth model is trained on an image set composed of all three strip types. The results from these models show it is able to predict positive and negative results to a high level of accuracy.

The field of biomedical research relies on the knowledge of binding interactions between various proteins of interest to create novel molecular targets for therapeutic purposes. While many of these interactions remain a mystery, knowledge of these properties and interactions could have significant medical applications in terms of understanding cell signaling and immunological defenses. Furthermore, there is evidence that machine learning and peptide microarrays can be used to make reliable predictions of where proteins could interact with each other without the definitive knowledge of the interactions. In this case, a neural network was used to predict the unknown binding interactions of TNFR2 onto LT-ɑ and TRAF2, and PD-L1 onto CD80, based off of the binding data from a sampling of protein-peptide interactions on a microarray. The accuracy and reliability of these predictions would rely on future research to confirm the interactions of these proteins, but the knowledge from these methods and predictions could have a future impact with regards to rational and structure-based drug design.