About

Emily Kyle Fox is a Teaching Associate Professor at the Siebel School of Computing and Data Science at the University of Illinois Urbana-Champaign. She completed her Ph.D. in Computer Science at the same university in 2013, following her M.S. in 2010 and B.S. in 2008, all from the University of Illinois Urbana-Champaign. Her research interests include algorithms, computational geometry and topology, and graph algorithms. Fox has contributed to the field through various publications in reputable journals and conference proceedings, focusing on topics such as clustering with faulty centers, minimum cuts in hypergraphs and surface graphs, geometric edit distance, and approximation schemes for geometric transportation. She has received recognition for her teaching, including the Best Teacher award at the University of Texas at Dallas for the 2019-2020 academic year, and her research has been supported by the National Science Foundation CAREER Award. Her academic career includes positions at the University of Texas at Dallas before joining the University of Illinois Urbana-Champaign.

Research topics

• Artificial Intelligence
• Computer Science
• Information Retrieval
• Natural Language Processing
• Data Mining
• Machine Learning
• Database
• Data science
• Linguistics
• Psychology
• Law

Selected publications

• From Data Deficient to Big Data in Shark Conservation

  Fish and Fisheries · 2025-08-11 · 1 citations

  articleOpen access

  ABSTRACT Citizen science is increasingly harnessed worldwide to gather data otherwise requiring a prohibitive investment of funding and time. Meanwhile, the revolution in digital communication offers opportunities from crowdsourcing, big data approaches and social network mining to quickly and cost‐effectively fill major gaps in knowledge necessary to protect endangered populations. Sharks are among the most endangered and data‐poor vertebrates in the ocean. Mainly due to overfishing, many shark populations are declining worldwide, while most species lack basic abundance, distribution and life‐history data. Hence, filling knowledge gaps across taxa, ecosystems, and regions is urgently needed to increase our understanding of their ecology, develop effective conservation actions and reverse their loss. Here, we introduce a novel citizen science and crowdsourcing approach for conservation through sharkPulse, a new platform automating data ingestion and organisation to build the largest database of shark occurrence records to date. Designed to complement and extend similar biodiversity monitoring tools relying heavily on user submissions, sharkPulse aims to source large streams of online shark images and transform them into occurrence records, filling knowledge gaps in shark ecology and biology. This platform offers a blueprint to leverage AI and big data approaches, social network data mining and participatory science to efficiently and continuously source visual media materials and transform the monitoring of data‐limited marine and terrestrial animal populations.

  PublisherOA PDFDOI

• AI-Facilitated Episodic Future Thinking For Adults with Obesity

  ArXiv.org · 2025-03-08

  preprintOpen accessSenior author

  Episodic Future Thinking (EFT) involves vividly imagining personal future events and experiences in detail. It has shown promise as an intervention to reduce delay discounting-the tendency to devalue delayed rewards in favor of immediate gratification- and to promote behavior change in a range of maladaptive health behaviors. We present EFTeacher, an AI chatbot powered by the GPT-4-Turbo large language model, designed to generate EFT cues for users with lifestyle-related conditions. To evaluate the feasibility and usability of EFTeacher, we conducted a mixed-methods study that included usability assessments, user evaluations based on content characteristics questionnaires, and semi-structured interviews. Qualitative findings indicate that participants perceived EFTeacher as communicative and supportive through an engaging dialogue. The chatbot facilitated imaginative thinking and reflection on future goals. Participants appreciated its adaptability and personalization features, though some noted challenges such as repetitive dialogue and verbose responses. Our findings underscore the potential of large language model-based chatbots in EFT interventions targeting maladaptive health behaviors.

  PublisherOA PDFDOI

• Don’t let food poisoning crash your picnic – six tips to keep your spread safe

  2025-07-11

  preprint1st authorCorresponding
  PublisherDOI

• What’s in a cue?: Using natural language processing to quantify content characteristics of episodic future thinking in the context of overweight and obesity

  Health Psychology and Behavioral Medicine · 2025-06-02

  articleOpen access

  Episodic future thinking (EFT), an intervention in which participants vividly imagine their future, has been explored as a cognitive intervention to reduce delay discounting and decrease engagement in harmful health behaviors. In these studies, participants generate text descriptions of personally meaningful future events. The content of these text descriptions, or cues, is heterogeneous and can vary along several dimensions (e.g. references to health, celebrations, family; vividness; emotional valence). However, little work has quantified this heterogeneity or potential importance for EFT's efficacy. To better understand the potential impact of EFT content in the context of health behavior change (e.g. diet) among people with or at risk for obesity and related conditions, we used data from 19 prior EFT studies, including 1705 participants (mean body mass index = 33.1) who generated 9714 cues. We used natural language processing to classify EFT content and examined whether EFT content moderated effects on delay discounting. Cues most commonly involved recreation, food, and spending time with family, and least commonly involved references to health and self-improvement. Cues were generally classified as highly vivid, episodic, and positively valent (consistent with the intervention's design). In multivariate regression with model selection, EFT content did not significantly moderate the effect of the episodic thinking intervention. Thus, we find no evidence that any of the content characteristics we examined were important moderators of the efficacy of EFT in reducing delay discounting. This suggests that EFT's efficacy is robust against variability in these characteristics. However, note that in all studies, EFT methods were designed to generate high levels of vividness, episodicity, and emotional valence, potentially resulting in a ceiling effect in these content areas. Moreover, EFT content was not experimentally manipulated, limiting causal inference. Future studies should experimentally examine these and other content characteristics and evaluate their possible role in EFT's efficacy.

  PublisherDOI

• Toward Robust URL Extraction for Open Science: A Study of arXiv File Formats and Temporal Trends

  ArXiv.org · 2025-09-05

  preprintOpen access

  In this work, we study how URL extraction results depend on input format. We compiled a pilot dataset by extracting URLs from 10 arXiv papers and used the same heuristic method to extract URLs from four formats derived from the PDF files or the source LaTeX files. We found that accurate and complete URL extraction from any single format or a combination of multiple formats is challenging, with the best F1-score of 0.71. Using the pilot dataset, we evaluate extraction performance across formats and show that structured formats like HTML and XML produce more accurate results than PDFs or Text. Combining multiple formats improves coverage, especially when targeting research-critical resources. We further apply URL extraction on two tasks, namely classifying URLs into open-access datasets and software and the others, and analyzing the trend of URLs usage in arXiv papers from 1992 to 2024. These results suggest that using a combination of multiple formats achieves better performance on URL extraction than a single format, and the number of URLs in arXiv papers has been steadily increasing since 1992 to 2014 and has been drastically increasing from 2014 to 2024. The dataset and the Jupyter notebooks used for the preliminary analysis are publicly available at https://github.com/lamps-lab/arxiv-urls

  PublisherOA PDFDOI

• Five Years of Ublituximab in Relapsing Multiple Sclerosis: Additional Results from Open-label Extension of ULTIMATE I and II Studies (P11-1.006)

  Neurology · 2025-04-07

  article

  To evaluate the long-term clinical efficacy and safety of ublituximab.

  PublisherDOI

• VTechAGP: An Academic-to-General-Audience Text Paraphrase Dataset and Benchmark Models

  2025-01-01

  articleOpen access

  Ming Cheng, Jiaying Gong, Chenhan Yuan, William A Ingram, Edward Fox, Hoda Eldardiry. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

  PublisherOA PDFDOI

• Analyzing Quality, Bias, and Performance in Text-to-Image Generative Models

  arXiv (Cornell University) · 2024-06-28 · 2 citations

  preprintOpen accessSenior author

  Advances in generative models have led to significant interest in image synthesis, demonstrating the ability to generate high-quality images for a diverse range of text prompts. Despite this progress, most studies ignore the presence of bias. In this paper, we examine several text-to-image models not only by qualitatively assessing their performance in generating accurate images of human faces, groups, and specified numbers of objects but also by presenting a social bias analysis. As expected, models with larger capacity generate higher-quality images. However, we also document the inherent gender or social biases these models possess, offering a more complete understanding of their impact and limitations.

  PublisherOA PDFDOI

• Multi-dimensional Edge-Embedded GCNs for Arabic Text Classification

  Lecture notes in computer science · 2024-01-01

  book-chapterSenior author
  PublisherDOI

• Automating Chapter-Level Classification for Electronic Theses and Dissertations

  2024-12-15 · 2 citations

  articleSenior author

  Traditional archival practices for describing electronic theses and dissertations (ETDs) rely on broad, high-level metadata schemes that fail to capture the depth, complexity, and interdisciplinary nature of these long scholarly works. The lack of detailed, chapter-level content descriptions impedes researchers’ ability to locate specific sections or themes, thereby reducing discoverability and overall accessibility. By providing chapter-level metadata information, we improve the effectiveness of ETDs as research resources. This makes it easier for scholars to navigate them efficiently and extract valuable insights. The absence of such metadata further obstructs interdisciplinary research by obscuring connections across fields, hindering new academic discoveries and collaboration. In this paper, we propose a machine learning and AI-driven solution to automatically categorize ETD chapters. This solution is intended to improve discoverability and promote understanding of chapters. Our approach enriches traditional archival practices by providing context-rich descriptions that facilitate targeted navigation and improved access. We aim to support interdisciplinary research and make ETDs more accessible. By providing chapter-level classification labels and using them to index in our developed prototype system, we make content in ETD chapters more discoverable and usable for a diverse range of scholarly needs. Implementing this AI-enhanced approach allows archives to serve researchers better, enabling efficient access to relevant information and supporting deeper engagement with ETDs. This will increase the impact of ETDs as research tools, foster interdisciplinary exploration, and reinforce the role of archives in scholarly communication within the data-intensive academic landscape.

  PublisherDOI

Recent grants

• SGER: DL-VT416: A Digital Library Testbed for Research Related to 4/16/2007 at Virginia Tech

  NSF · $212k · 2007–2009

• Collaborative Project: Ensemble: Enriching Communities and Collections to Support Education in Computing

  NSF · $510k · 2008–2014

• III: Small: Integrated Digital Event Archiving and Library (IDEAL)

  NSF · $500k · 2013–2017

• I-Corps: Automated Summarization Technology

  NSF · $50k · 2019–2021

• III:Small:Integrated Digital Library Support for Crisis, Tragedy, and Recovery

  NSF · $500k · 2009–2013

Frequent coauthors

• Krzysztof Selmaj

  University of Warmia and Mazury in Olsztyn

  252 shared

• Robert J. Fox

  245 shared

• J. Theodore Phillips

  196 shared

• Eva Havrdová

  General University Hospital in Prague

  175 shared

• Michael Hutchinson

  University College Dublin

  161 shared

• Mariko Kita

  Virginia Mason Medical Center

  161 shared

• Robert Herndon

  California State Polytechnic University

  147 shared

• Ralf Gold

  Ruhr University Bochum

  147 shared

Education

• Ph.D., Computer Science

  University of Illinois at Urbana-Champaign

  2005

• M.S., Computer Science

  University of Illinois at Urbana-Champaign

  2001

• B.S., Computer Science

  University of Illinois at Urbana-Champaign

  1999

Awards & honors

• Best Teacher in Computer Science, Erik Jonsson School of Eng…
• National Science Foundation CAREER Award (CCF-1942597) (2020…