A. James Clark School of Engineering
Permanent URI for this communityhttp://hdl.handle.net/1903/1654
The collections in this community comprise faculty research works, as well as graduate theses and dissertations.
Browse
3 results
Search Results
Item EFFECTS OF DIVERSE INITIALIZATION ON BAYESIAN OPTIMIZERS(2023) Kamrah, Eesh; Fuge, Mark D; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Design researchers have struggled to produce quantitative predictions for exactly why andwhen diversity might help or hinder design search efforts. This thesis addresses that problem by studying one ubiquitously used search strategy—Bayesian Optimization (BO)—on different ND test problems with modifiable convexity and difficulty. Specifically, we test how providing diverse versus non-diverse initial samples to BO affects its performance during search and introduce a fast ranked-DPP method for computing diverse sets, which we need to detect sets of highly diverse or non-diverse initial samples. We initially found, to our surprise, that diversity did not appear to affect BO, neither helping nor hurting the optimizer’s convergence. However, follow-on experiments illuminated a key trade-off. Non-diverse initial samples hastened posterior convergence for the underlying model hyper-parameters—a Model Building advantage. In contrast, diverse initial samples accelerated exploring the function itself—a Space Exploration advantage. Both advantages help BO, but in different ways, and the initial sample diversity directly modulates how BO trades those advantages. Indeed, we show that fixing the BO hyper-parameters removes the Model Building advantage, causing diverse initial samples to always outperform models trained with non-diverse samples. These findings shed light on why, at least for BO-type optimizers, the use of diversity has mixed effects and cautions against the ubiquitous use of space-filling initializations in BO. To the extent that humans use explore-exploit search strategies similar to BO, our results provide a testable conjecture for why and when diversity may affect human-subject or design team experiments. The thesis is organized as follows: Chapter 2 provides an overview of existing studies that explore the impact of different initial stimuli. In Chapter 3, we explain the methodology used in the subsequent experiments. Chapter 4 presents the results of our initial study on the diverse initialization of BO (Bayesian Optimization) applied to the wildcat wells function. In this chapter we also investigate the conditions under which less diverse initial examples perform better and expand on these findings in Chapter 5 by considering additional ND continuous functions. The final chapter discusses the limitations of our findings and proposes potential areas for future research.Item Root Cause Analysis Of Adverse Events Using A Human Reliability Analysis Approach(2022) Johnson, David Michael; Vaughn-Cooke, Monifa; Reliability Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Large scale analysis of adverse event data is challenging due to the unstructured nature of event reporting and narrative textual data in adverse event repositories. This issue is further complicated for human error adverse events, which are routinely treated as a root cause instead of as initiating events in a causal chain. Human error events are commonly misunderstood and underreported, which hinders the analysis of trends and the identification of risk mitigation strategies across industries. Currently, the prevailing means of human error investigation is the analysis of accident and incident data which are not designed around a framework of human cognition or psychomotor function. Existing approaches lack a theoretical foundation with sufficient cognitive granularity to identify root causes of human error. This research provides a cognitive task decomposition to standardize the investigation, reporting, and analysis of human error adverse event data in narrative textual form. The proposed method includes a qualitative structure to answer six questions (when, who, what, where, how, why) that are critical to comprehensively understand the events surrounding human error. This process is accomplished in five main stages: (1) Develop guidelines for a cognitively-driven adverse event investigation; (2) Perform a baseline cognitive task analysis (when) to document relevant stakeholders (who), products or processes (what), and environments (where) based on a taxonomy of cognitive and psychomotor function; (3) Identify deviations for the baseline task analysis in the form of unsafe acts (how) using a human error classification; (4) and Develop a root cause mapping to identify the performance shaping factors (PSFs) (why) for each unsafe act. The outcome of the proposed method will advance the fields of risk analysis and regulatory science by providing a standardized and repeatable process to input and analyze human error in adverse event databases. The method provides a foundation for more effective human error trending and accident analysis at a greater level of cognitive granularity. Application of this method to adverse event risk mitigations can inform prospective strategies such as resource allocation and system design, with the ultimate long-term goal of reducing the human contribution to risk.Item Fostering Creativity in Engineering Students Using Psychological Theory(2018) Leonard, Jared; Schmidt, Linda C.; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Four experiments were conducted exploring the application of modern psychological theory to improving creative performance in engineering students, as measured by the divergent thinking test the Alternative Uses Task [AUT] and the graphical data analysis method linkography. Evidence was found for the presence of the serial order effect, but not for the efficacy of incubation or direct instruction in the psychology of creativity. A more practical test and instruction may be required. Making a meaningful improvement in the creativity of engineering students may require broad, systemic change in the way engineering is taught.