Matching Items (5)
Filtering by
- All Subjects: Machine Learning
- Resource Type: Text
Description
This paper seeks to analyze the relationship between energy subsidies on fossil fuels by countries and corresponding energy consumption, specifically electricity, by its citizens and occupants. The purpose of this was to determine whether pre-tax subsidies and post-tax subsidies have an effect on that consumption. This paper will discuss the prospect of accounting for post-tax subsidies as a method to curb rampant energy consumption throughout the world, with the focus being on residential electricity use. The two case studies, the Netherlands and Saudi Arabia, will illustrate the consumption patterns in relatively similar economic societies with different subsidy policies. Saudi Arabia will be a high pre-tax subsidy example while the Netherlands will be shown to account for some of the post-tax subsidies through an externality tax system. At the end of this analysis, this paper will show that the heavy subsidization of electricity production is strongly correlated to residential electricity consumption at levels that many officials would deem unsustainable, and that as such, subsidy reform is both beneficial and necessary.
ContributorsCorona, Kyle (Author) / Kelman, Jonathan (Thesis director) / Breetz, Hanna (Committee member) / School of Sustainability (Contributor, Contributor) / Economics Program in CLAS (Contributor) / Barrett, The Honors College (Contributor)
Created2016-12
Description
2018, Google researchers published the BERT (Bidirectional Encoder Representations from Transformers) model, which has since served as a starting point for hundreds of NLP (Natural Language Processing) related experiments and other derivative models. BERT was trained on masked-language modelling (sentence prediction) but its capabilities extend to more common NLP tasks, such as language inference and text classification. Naralytics is a company that seeks to use natural language in order to be able to categorize users who create text into multiple categories – which is a modified version of classification. However, the text that Naralytics seeks to pull from exceed the maximum token length of 512 tokens that BERT supports – so this report discusses the research towards multiple BERT derivatives that seek to address this problem – and then implements a solution that addresses the multiple concerns that are attached to this kind of model.
ContributorsNgo, Nicholas (Author) / Carter, Lynn (Thesis director) / Lee, Gyou-Re (Committee member) / Barrett, The Honors College (Contributor) / Computer Science and Engineering Program (Contributor) / Economics Program in CLAS (Contributor)
Created2023-05
Description
A defense-by-randomization framework is proposed as an effective defense mechanism against different types of adversarial attacks on neural networks. Experiments were conducted by selecting a combination of differently constructed image classification neural networks to observe which combinations applied to this framework were most effective in maximizing classification accuracy. Furthermore, the reasons why particular combinations were more effective than others is explored.
ContributorsMazboudi, Yassine Ahmad (Author) / Yang, Yezhou (Thesis director) / Ren, Yi (Committee member) / School of Mathematical and Statistical Sciences (Contributor) / Economics Program in CLAS (Contributor) / Barrett, The Honors College (Contributor)
Created2019-05
Description
This project seeks to provide a general picture of the economic dependence on fossil fuels per County in the United States. The purpose for this study is creating a foundation for conversations about the future of fossil fuel workers and counties that depend heavily on fossil fuels. The main indicators utilized for this were employment and payroll data extracted from United States Census Bureau’s County Business Patterns dataset. A section on similarities between fossil fuel workers and other occupations was included, which shows possible alternative industries for fossil fuel workers. The main goal of the project is to provide possible solutions for mitigating job losses in the future. Some proposed solutions include retraining, expanding higher education, and investing in new industries. It is most important for future work to include input from most vulnerable counties and understand the social and cultural complexities that are tied to this problem.
ContributorsRamirez Torres, Jairo Adriel (Author) / Miller, Claek (Thesis director) / Shutters, Shade (Committee member) / Watts College of Public Service & Community Solut (Contributor) / Electrical Engineering Program (Contributor) / Economics Program in CLAS (Contributor) / Barrett, The Honors College (Contributor)
Created2020-05
Using Machine Learning Classification Techniques to Predict Recessionary Periods in the U.S. Economy
Description
The goal of this research project is to determine how beneficial machine learning (ML) techniquescan be in predicting recessions. Past work has utilized a multitude of classification methods from Probit models to linear Support Vector Machines (SVMs) and obtained accuracies nearing 60-70%, where some models even predicted the Great Recession based off data from the previous 50 years. This paper will build on past work, by starting with less complex classification techniques that are more broadly used in recession forecasting and end by incorporating more complex ML models that produce higher accuracies than their more primitive counterparts. Many models were tested in this analysis and the findings here corroborate past work that the SVM methodology produces more accurate results than currently used probit models, but adds on that other ML models produced sufficient accuracy as well.
ContributorsHogan, Carter (Author) / McCulloch, Robert (Thesis director) / Pereira, Claudiney (Committee member) / Barrett, The Honors College (Contributor) / School of International Letters and Cultures (Contributor) / Economics Program in CLAS (Contributor) / School of Mathematical and Statistical Sciences (Contributor)
Created2022-05