SXSean Xin Xu

Sean Xin Xu

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Sean Xin Xu has a total of 54 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Bounded Rationality in the Digital Age: How Salient Negative Secondhand Information on Knowledge-Sharing Platforms Triggers Investor Mis-Reactions

    by Zhitao Yin, Ning Jia, Sean Xin Xu, Xiaoquan (Michael) Zhang

    Digital platforms increasingly shape how retail investors encounter firm information, not only by spreading new or false content but also by resurfacing accurate information that was already public. This study examines whether such secondhand information affects investor behavior when it becomes salient and negative. In a randomized field experiment on Wikipedia, we added previously disclosed litigation news to selected public firms’ pages and compared outcomes with matched control firms and a placebo group whose pages received routine operations news. The litigation news additions increased Wikipedia pageviews, raised retail trading volume by 15.7%, and narrowed bid-ask spreads by 6.9 basis points; the placebo edits did not produce comparable effects. The effects were more evident when litigation was severe or recent, when firms were more visible or had more positive prior sentiment, and when the litigation signaled governance weakness. These findings suggest that already public negative information can still shape market behavior when platform-mediated resurfacing draws attention to it. Regulators can strengthen investor education about reshared content. Firms can monitor digital platforms for resurfaced sensitive information. Platforms can provide clearer disclosure date and novelty cues. Financial professionals can incorporate source and novelty checks before acting on attention-grabbing firm information.

  • Top Cited Papers
    Open Access

    Human-AI Collaborative Essay Scoring: A Dual-Process Framework with LLMs

    by Changrong Xiao, Wenxing Ma, Qingping Song, Sean Xin Xu, Kunpeng Zhang, Yufang Wang, Qi Fu


  • Open Access

    Enhancing Investment Analysis: Optimizing AI-Agent Collaboration in Financial Research

    by Xuewen Han, Neng Wang, Shangkun Che, Hongyang Yang, Kunpeng Zhang, Sean Xin Xu

    In recent years, the application of generative artificial intelligence (GenAI) in financial analysis and investment decision-making has gained significant attention. However, most existing approaches rely on single-agent systems, which fail to fully utilize the collaborative potential of multiple AI agents. In this paper, we propose a novel multi-agent collaboration system designed to enhance decision-making in financial investment research. The system incorporates agent groups with both configurable group sizes and collaboration structures to leverage the strengths of each agent group type. By utilizing a sub-optimal combination strategy, the system dynamically adapts to varying market conditions and investment scenarios, optimizing performance across different tasks. We focus on three sub-tasks: fundamentals, market sentiment, and risk analysis, by analyzing the 2023 SEC 10-K forms of 30 companies listed on the Dow Jones Index. Our findings reveal significant performance variations based on the configurations of AI agents for different tasks. The results demonstrate that our multi-agent collaboration system outperforms traditional single-agent models, offering improved accuracy, efficiency, and adaptability in complex financial environments. This study highlights the potential of multi-agent systems in transforming financial analysis and investment decision-making by integrating diverse analytical perspectives.


  • On Crafting Effective Theoretical Contributions for Empirical Papers in Economics of Information Systems: Some Editorial Reflections

    by Anandasivam Gopal, Pei-Yu Chen, Wonseok Oh, Sean Xin Xu, Suprateek Sarker

    The terms theory and theoretical contributions evoke mixed reactions in the information systems discipline, especially among empirical researchers in the economics of information systems (Econ-IS) area. Although some see such contributions as the raison d’etre for academic scholars engaged in research, others feel that the discipline has developed a fetish for theory, with reviewers and editors often demanding an unreasonable level of theoretical contributions for empirical manuscripts to succeed in the review process. Moreover, there exists a great deal of diversity in the conception of what constitutes a reasonable theoretical contribution, especially within empirical work, across editors and reviewers, leading to frustration with the review process and disappointment with editorial decisions. Given the different types of theoretical contributions that may be suitable for a given manuscript and recognizing the changing nature of empirical work within Econ-IS, we attempt to shed some light on theoretical contributions within empirical Econ-IS research, paying attention to their nature, types, and impact. Specifically, we start by reflecting on the typical theory-related comments we have seen in review packets that we generalize to a set of critiques often related to empirical papers. Subsequently, we provide a working definition of a theoretical contribution and the components that make up such a contribution. We then propose a taxonomy of theoretical contributions typically observed in Information Systems Research (ISR). Based on this taxonomy of contributions, the typical critiques observed in empirical Econ-IS papers, and a set of published papers, we provide some broad guidelines for how authors may craft an effective theoretical contribution for submission to ISR. We also discuss a pathway for manuscripts that do not (seek to) offer significant theoretical contributions. Such manuscripts are welcome, but we believe that a very high bar of practical impact must be met for them to succeed in the review process. Based on the guidelines and suggestions made here, our hope is that authors and evaluators will participate in the review process with a shared understanding of the elusive notion of theoretical contributions.


  • Open Access

    Multimodal Data Augmentation for Image Captioning using Diffusion Models

    by Changrong Xiao, Sean Xin Xu, Kunpeng Zhang

    Image captioning, an important vision-language task, often requires a tremendous number of finely labeled image-caption pairs for learning the underlying alignment between images and texts. In this paper, we proposed a multimodal data augmentation method, leveraging a recent text-to-image model called Stable Diffusion, to expand the training set via high-quality generation of image-caption pairs. Extensive experiments on the MS COCO dataset demonstrate the advantages of our approach over several benchmark methods, and particularly a significant boost when having fewer training instances. In addition, models trained on our augmented datasets also outperform prior unpaired image captioning methods by a large margin. Finally, further improvement regarding the training efficiency and effectiveness can be obtained after intentionally filtering the generated data based on quality assessment.


  • Open Access

    Multimodal Data Augmentation for Image Captioning using Diffusion Models

    by Changrong Xiao, Sean Xin Xu, Kunpeng Zhang

    Image captioning, an important vision-language task, often requires a tremendous number of finely labeled image-caption pairs for learning the underlying alignment between images and texts. In this paper, we proposed a multimodal data augmentation method, leveraging a recent text-to-image model called Stable Diffusion, to expand the training set via high-quality generation of image-caption pairs. Extensive experiments on the MS COCO dataset demonstrate the advantages of our approach over several benchmark methods, and particularly a significant boost when having fewer training instances. In addition, models trained on our augmented datasets also outperform prior unpaired image captioning methods by a large margin. Finally, further improvement regarding the training efficiency and effectiveness can be obtained after intentionally filtering the generated data based on quality assessment.


  • Top Cited Papers
    Open Access

    Evaluating Reading Comprehension Exercises Generated by LLMs: A Showcase of ChatGPT in Education Applications

    by Changrong Xiao, Sean Xin Xu, Kunpeng Zhang, Yufang Wang, Lei Xia

    Changrong Xiao, Sean Xin Xu, Kunpeng Zhang, Yufang Wang, Lei Xia. Proceedings of the 18th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2023). 2023.


  • Configuring the Enterprise Systems Portfolio: The Role of Information Risk

    by Chaitanya Sambhara, Arun Rai, Sean Xin Xu

    Information risk, the likelihood that corporate financial information is of poor quality, adversely impacts investor confidence regarding a firm’s financial health, making it an economically important problem. Viewing a firm’s enterprise systems (ES) portfolio as made up of operational modules (customer relationship management and supply chain management) and functional modules (accounting and finance, and human resource management), we examine how firms configure their ES portfolio by changing the balance in the implementation of two types of modules in response to information risk. We find internal controls to be an important contingency in determining how firms change their ES portfolio balance when information risk increases. When there is no weakness in internal controls, firms change their ES portfolio balance more toward operational modules. However, when internal controls are afflicted with material weakness, firms change their ES portfolio balance more toward functional modules instead. When evaluating the link between ES portfolio configuration and information processing requirements in the context of financial processes, managers should assess both information risk and internal controls to decide how to change the balance between operational and functional modules that are implemented.


  • Examining Assumptions: Provocations on the Nature, Impact, and Implications of IS Theory

    by Andrew Burton-Jones, Brian S. Butler, Susan V. Scott, Sean Xin Xu

    The Information Systems research community has a complex relationship with theory and theorizing. As a community of scholars, our assumptions about theory and theorizing affect every aspect of our intellectual lives. Ideas about what theory is, who theorizes, where theory comes from, when we theorize, how theory is developed and changes, and why theory is (or isn’t) important shapes the projects we do, the partnerships we have, the resources available to us, and phenomena that we find to be significant, interesting, and novel. Dspite its prominent place in our thinking, it is often difficult to critically examine the assumptions we make about theory and theorizing. Our assumptions affect our priorities, decision, and actions. How we think about theory and its role in our individual and collective intellectual lives is a product of complex paths and multiple influences. Although we may try to stay mindful of how our assumptions shape our perspectives and ways of thinking, from time to time we need to look beyond them. The short papers presented here highlight and challenge some core assumptions. Each provocatively engages with one or more assumptions about theory, theorizing, and/or their implications for IS research. In doing so, these authors engage and contribute to the long tradition of reflecting on theory and theorizing. They challenge us to return to these dialogues anew as the worlds we study and the ones we inhabit change. Readers will also find healthy differences in opinion across the provocations. We expect the diversity in these provocations will inspire thought and reflection as much as the strong views within each one.


  • Top Cited Papers

    How Firms Make Information Technology Investment Decisions: Toward a Behavioral Agency Theory

    by John Qi Dong, Prasanna P. Karhade, Arun Rai, Sean Xin Xu

    While prior research has established that information technology (IT) investment has a significant impact on firm performance, relatively few studies have provided insights into the antecedents of IT investment decisions. By integrating the behavioral theory of the firm and agency theory, we propose a behavioral agency theory to explain performance shortfalls and corporate governance, which monitors and controls managers’ tendency of overinvestment or underinvestment in IT, as key drivers that jointly determine IT investment. As such, IT investment facilitates a firm’s problemistic search that generates innovation in response to performance gaps. We further examine the role of innovation outputs as a mediating mechanism linking IT investment to firm performance. Our econometric analysis of a large-scale panel dataset provides empirical evidence corroborating our theory. Overall, this study contributes a behavioral agency theory to deepen our understanding about performance drivers and outcomes of IT investment decisions.