About

Hoda Eldardiry is an Associate Professor and the Director of the Machine Learning Laboratory at Virginia Tech. She specializes in artificial intelligence, machine learning, data mining, statistical relational learning, and social network analysis. Her research focuses on developing advanced algorithms and models to analyze complex data structures and social networks, contributing to the fields of data science and computational intelligence. She holds a Ph.D. and M.S. in computer science from Purdue University and a B.E. in computer and systems engineering from Alexandria University in Egypt. Her academic and research activities are based at Virginia Tech's Data and Decision Sciences Building in Blacksburg, VA, and she is involved in various research initiatives and collaborations related to her expertise.

Research topics

• Computer Science
• Artificial Intelligence
• Machine Learning
• Economics
• Econometrics
• Automotive engineering
• Environmental economics
• Finance
• Transport engineering
• Environmental science
• Engineering
• Mathematics
• Theoretical computer science
• Statistics

Selected publications

• Framing Discussions of AI Policy Implications in Computing Courses

  2026-02-13

  articleOpen accessSenior author

  The growth and permeation of artificial intelligence (AI) technologies across society has drawn focus to the ways in which the responsible use of these technologies can be facilitated through AI governance. Increasingly, large companies and governments alike have begun to articulate and, in some cases, enforce governance preferences through AI policy. In this context, overlapping jurisdictions and even contradictory policy preferences across private companies, local, national, and multinational governments create a complex landscape for AI policy which, we argue, will require AI developers able adapt to an evolving regulatory environment. Preparing CS students for the new challenges of an AI-saturated technology industry should therefore constitute a key priority for the computing curriculum. In this work, we will outline a proposed framework for integrating discussions on the nascent AI policy landscape into computer science courses. Building on recent literature on AI governance and our synthesis of AI policy efforts in the United States and European Union, we propose guiding questions to frame class discussions around AI policy in technical and non-technical (e.g., ethics) CS courses. Throughout, we emphasize the connection between normative policy demands and still-open technical challenges relating to their implementation and enforcement through code and governance structures. We conclude by highlighting opportunities to utilize our framework in practice, reflecting on our experiences using this framework in piloting curricular interventions in the 2024-2025 and 2025-2026 academic years.

  PublisherDOI

• Developing a Typology of Roles for STEM-Trained Professionals in AI Policy Engagement

  Bulletin of Science Technology & Society · 2026-01-09

  article

  This paper explores the roles of STEM (science, technology, engineering, and mathematics) professionals in AI policymaking, addressing the urgent need for informed governance in emerging technologies. With AI's complex sociotechnical impacts, STEM expertise is crucial for balancing benefits and mitigating risks like bias and privacy concerns. Despite their potential influence, the specific contributions of STEM professionals in AI policy remain underexplored. To address this gap, semi-structured interviews were conducted with 15 STEM professionals who have both educational and professional experience in STEM and policy, and have actively participated in AI policymaking. The data revealed two primary role groups: (1) advisors, gatekeepers, and influencers, who leverage their expertise in STEM and policy to guide stakeholders and actively influence policy decisions, and (2) facilitators and brokers, who facilitate connections and manage the flow of information between stakeholders. This typology highlights the varied contributions of STEM professionals to AI governance and policy development.

  PublisherDOI

• Alleviating Community Fear in Disasters via Multi-Agent Actor-Critic Reinforcement Learning

  arXiv (Cornell University) · 2026-04-09

  preprintOpen accessSenior author

  During disasters, cascading failures across power grids, communication networks, and social behavior amplify community fear and undermine cooperation. Existing cyber-physical-social (CPS) models simulate these coupled dynamics but lack mechanisms for active intervention. We extend the CPS resilience model of Valinejad and Mili (2023) with control channels for three agencies, communication, power, and emergency management, and formulate the resulting system as a three-player non-zero-sum differential game solved via online actor-critic reinforcement learning. Simulations based on Hurricane Harvey data show 70% mean fear reduction with improved infrastructure recovery; cross-validation in the case of Hurricane Irma (without refitting) achieves 50% fear reduction, confirming generalizability.

  PublisherDOI

• EQUAL: Entity-Enhanced QUery Expansion for EquitAble Crisis Summarization via KnowLedge Graphs

  Proceedings of the ... International ISCRAM Conference · 2026-05-22

  articleOpen access

  Disaster response efforts face persistent challenges in ensuring equitable aid and information access for all affected populations. Marginalized communities, including the elderly, persons with disabilities, people experiencing homelessness, low-income households, non-English speakers, and geographically isolated residents, face heightened risk during disasters and are more likely to experience delays in receiving aid, evacuation support, and critical information (Wilson et al. 2021). We present EQUAL (entity-enhanced query expansion for equitable crisis summarization via knowledge graphs), a work-in-progress framework that combines dual-model entity-enhanced query expansion with an equity-aware GraphRAG (Graph-based Retrieval-Augmented Generation) pipeline. EQUAL constructs crisis knowledge graphs enriched with vulnerability–resource connections and generates summaries through community-level synthesis. Evaluated on 18 real-world disaster events from the TREC CrisisFACTS dataset, EQUAL outperforms all baselines on equity-focused metrics—vulnerable group coverage, intersectional coverage, and statistical parity—and shows marked gains in explicit mentions of vulnerable populations, geographic specificity, and actionable resource information. It also remains competitive on standard semantic quality metrics.

  PublisherOA PDFDOI

• Scoping the AI curriculum: key competencies for future AI practitioners

  AI and Ethics · 2026-02-25

  articleOpen accessSenior authorCorresponding

  The sustained interest in artificial intelligence (AI) as an area of postsecondary study is evident in growing enrollment figures and in expanded course offerings focusing on subjects like machine learning, natural language processing, and computer vision. Yet, for the most part, the contributions of these courses to the development of future AI practitioners is considered only in isolation. We find that a comprehensive framework to conceptualize the combined impact of these courses throughout a student’s college education is lacking. Building on our previous research–particularly a study of computer science (CS) student attitudes and competencies related to AI and AI ethics–in this paper we begin to apply key findings towards the conceptualization of an AI curriculum. We argue that this curriculum must rest on three content pillars: technical foundations of AI systems, social implications of AI applications, and effective uses of AI tools. Such a holistic approach to AI education is necessary to underscore the centrality of policy considerations to the alignment of AI systems with normative objectives, and to reinforce an efficient and appropriate incorporation of AI into students’ productivity workflows. In an effort to imagine how the AI curriculum can better prepare the future AI workforce, we emphasize three core AI-related competencies that are currently under-developed among CS students studying AI. Then, we map each competency onto one or more skills which we argue should be fostered throughout the AI curriculum. Finally, we propose examples of curricular interventions to address each of these skills and provide one example of an AI-focused undergraduate course sequence, illustrating opportunities to construct a cohesive AI curriculum across multiple separate computing courses. In taking this ‘curriculum-level’ perspective, we proffer that our synthesis of disparate strands of inquiry through this paper constitutes an important contribution to the literature regarding the teaching about and teaching with AI.

  PublisherOA PDFDOI

• Alleviating Community Fear in Disasters via Multi-Agent Actor-Critic Reinforcement Learning

  arXiv (Cornell University) · 2026-04-09

  articleOpen accessSenior author

  During disasters, cascading failures across power grids, communication networks, and social behavior amplify community fear and undermine cooperation. Existing cyber-physical-social (CPS) models simulate these coupled dynamics but lack mechanisms for active intervention. We extend the CPS resilience model of Valinejad and Mili (2023) with control channels for three agencies, communication, power, and emergency management, and formulate the resulting system as a three-player non-zero-sum differential game solved via online actor-critic reinforcement learning. Simulations based on Hurricane Harvey data show 70% mean fear reduction with improved infrastructure recovery; cross-validation in the case of Hurricane Irma (without refitting) achieves 50% fear reduction, confirming generalizability.

  PublisherOA PDF

• InfinityStory: Unlimited Video Generation with World Consistency and Character-Aware Shot Transitions

  arXiv (Cornell University) · 2026-03-04

  articleOpen access

  Generating long-form storytelling videos with consistent visual narratives remains a significant challenge in video synthesis. We present a novel framework, dataset, and a model that address three critical limitations: background consistency across shots, seamless multi-subject shot-to-shot transitions, and scalability to hour-long narratives. Our approach introduces a background-consistent generation pipeline that maintains visual coherence across scenes while preserving character identity and spatial relationships. We further propose a transition-aware video synthesis module that generates smooth shot transitions for complex scenarios involving multiple subjects entering or exiting frames, going beyond the single-subject limitations of prior work. To support this, we contribute with a synthetic dataset of 10,000 multi-subject transition sequences covering underrepresented dynamic scene compositions. On VBench, InfinityStory achieves the highest Background Consistency (88.94), highest Subject Consistency (82.11), and the best overall average rank (2.80), showing improved stability, smoother transitions, and better temporal coherence.

  PublisherOA PDF

• InfinityStory: Unlimited Video Generation with World Consistency and Character-Aware Shot Transitions

  Open MIND · 2026-03-04

  preprint

  Generating long-form storytelling videos with consistent visual narratives remains a significant challenge in video synthesis. We present a novel framework, dataset, and a model that address three critical limitations: background consistency across shots, seamless multi-subject shot-to-shot transitions, and scalability to hour-long narratives. Our approach introduces a background-consistent generation pipeline that maintains visual coherence across scenes while preserving character identity and spatial relationships. We further propose a transition-aware video synthesis module that generates smooth shot transitions for complex scenarios involving multiple subjects entering or exiting frames, going beyond the single-subject limitations of prior work. To support this, we contribute with a synthetic dataset of 10,000 multi-subject transition sequences covering underrepresented dynamic scene compositions. On VBench, InfinityStory achieves the highest Background Consistency (88.94), highest Subject Consistency (82.11), and the best overall average rank (2.80), showing improved stability, smoother transitions, and better temporal coherence.

  DOI

• Use of machine learning algorithms to predict optimal hospital length of stay

  VTechWorks (Virginia Tech) · 2025-12-09

  articleSenior author

  Problem: Hospitals often struggle to allocate beds, equipment, and staff efficiently, leading to unnecessary complications. Predicting a patient’s length of stay (LOS) early helps hospitals plan treatment, staffing, and bed availability more effectively. Both extremes of LOS carry risks: discharging too early can result in inadequate care and higher readmissions, while prolonged stays waste resources and increase costs. Solution: Optimizing LOS improves patient outcomes using machine learning, enhances operational efficiency, and reduces overall spending.

  Publisher

• The AI Policy Module: Developing Computer Science Student Competency in AI Ethics and Policy

  2025-11-02

  articleSenior author
  PublisherDOI

Frequent coauthors

• Ryan A. Rossi

  23 shared

• Jiaying Gong

  12 shared

• Almuatazbellah Boker

  Virginia Tech

  11 shared

• Rong Zhou

  Shenzhen University

  9 shared

• Nesreen K. Ahmed

  Intel (United States)

  9 shared

• Hongjie Chen

  8 shared

• Kanak Mahadik

  8 shared

• Chenhan Yuan

  7 shared