Date of Award

2024-08-01

Degree Name

Doctor of Philosophy

Department

Computer Science

Advisor(s)

Natalia Villanueva Rosales

Abstract

The outcomes of scientific models can be hard to understand given the need for context, domain knowledge, and many variables and data being used. This research aims to provide users of scientific models with understandable information that can be automatically generated. For the context of this research, understandable is defined as being correctly interpreted (i.e., with respect to the original intent of the data). Scientific information can be conveyed to users in the form of a narrative or visualizations, and these are not necessarily separate from each other, but rather complimentary. One of the objectives of this research is to find ways to complement visual data with textual information to inform the user. In particular, this research aims to automate the creation of scientific narrative elements and the introduction of linguistic user profiles. Scientific narrative elements are natural language explanations aiming to provide context to results generated by scientific models, and semantic linguistic user profiles are containers of linguistic information for specific user groups in a specific context. The goal of this project is to foster the interpretability of scientific model outputs through the generation of narrative elements (i.e., natural language descriptions) that consider the context implicitly defined in the models. The project presents different approaches to achieve this goal, including meta-templates, grammatical frameworks, and linguistic profiles to automatically create scientific narrative elements. The SWIM (Sustainable Water through Integrated Modeling) framework is used as a case study to assess the creation of narrative elements for scientific water models. This work is evaluated through a comparison of the different approaches and an evaluation of the quality of narratives generated for different groups in two languages (i.e., English and Spanish). Results indicate that meta-templates and grammatical frameworks can be used to generate scientific narrative elements without the need of extensive training data. Additionally, linguistic profiles in the form of Grammatical Frameworks can be used to generate more stylistic scientific narrative elements for specific user groups. This research contributes to bridging a gap in contemporary scientific research by focusing on the automated generation of computational narratives, specifically for scientific models, that clarify the context, inputs, outputs, and potential scenarios of intricate scientific models. Future work includes the automation of linguistic profile Grammatical Framework linearization.

Language

en

Provenance

Received from ProQuest

File Size

129 p.

File Format

application/pdf

Rights Holder

Angel Uriel Ortega Castillo

Download

Share

COinS