Digital Commons Network^™

Metarag: Identifying Website Owner Using Meta-Path-Guided Dynamic Graph Retrieval-Augmented Generation, Cheng Tu, Yunshan Ma, Bingyang Guo, Qianyu Li, Yang Li, Min Zhang, Fan Shi, Xiang Wang

Research Collection School Of Computing and Information Systems

Website owner identification aims to link websites to their real-world owners, which is crucial for credibility assessment and information provenance in information retrieval and vital for applications in cybersecurity, Internet governance, and digital regulation. Existing approaches for website owner identification primarily rely on querying infrastructure registration records or analyzing webpage content. However, these methods often fail due to incomplete or outdated registration records and sparse webpage content. We observe that inter-website relationships, derived from shared infrastructure data such as primary domains, IP blocks, and geolocations, can provide valuable but underutilized ownership cues. To exploit this insight, we propose MetaRAG, a …

Stprompt++: Prompting Vision-Language Models For Weakly Supervised Video Anomaly Detection And Fine-Grained Localization, Peng Wu, Chengyu Pan, Guansong Pang, Xiangteng He, Zhiwei Yang, Peng Wang, Yanning Zhang

Research Collection School Of Computing and Information Systems

Traditional weakly supervised video anomaly detection (WSVAD) tasks typically rely on coarse-grained frame-level labels for training. Although this approach reduces annotation costs, it results in weak semantic understanding and spatial localization capabilities due to the absence of fine-grained annotations, hindering precise pixel-level anomaly detection and localization. Thanks to the success of vision-language models (VLMs), e.g., CLIP, recent approaches leveraging large VLMs focus on exploiting their strong semantic understanding capabilities, but they typically feed only keyframes or short video segments into the models, without supplying sufficient prior contextual information (e.g., contextual frames around anomalies, zoomed-in anomaly regions, and detailed anomaly descriptions), …

Ai Failures In The Eyes Of The Downstream Developer: A First Look At Concerns, Practices, And Challenges, Haoyu Gao, Mansooreh Zahedi, Wenxin Jiang, Hong Yi Lin, James C. Davis, Christoph Treude

Research Collection School Of Computing and Information Systems

With the advancement of AI models, more software systems are adopting AI as a component to facilitate automation. Pre-trained models (PTMs) have become a cornerstone of AI-based software, allowing for rapid integration and development with lower training cost. However, their adoption also introduces failure modes such as data leakage and biased outputs, that may require careful handling by downstream developers. While previous research has proposed taxonomies of these technical concerns and various mitigation strategies, how downstream developers address these issues during the development of general AI-based software when reusing PTMs remains unexplored. Understanding downstream developers’ perspectives is essential because they …

Logupdater: Automated Detection And Repair Of Specific Defects In Logging Statements, Renyi Zhong, Yichen Li, Jinxi Kuang, Wenwei Gu, Yintong Huo, Michael R. Lyu

Research Collection School Of Computing and Information Systems

Developers write logging statements to monitor software runtime behaviors and system state. However, poorly constructed or misleading log messages can inadvertently obfuscate actual program execution patterns, thereby impeding effective software maintenance. Existing research on analyzing issues within logging statements is limited, primarily focusing on detecting a singular type of defect and relying on manual intervention for fixes rather than automated solutions.To address the limitation, we initiate a systematic study that pinpoints four specific types of defects in logging statements (i.e., statement code inconsistency, static dynamic inconsistency, temporal relation inconsistency, and readability issues) through the analysis of real-world log-centric changes. We …

Generalizing Unsupervised Lidar Motion Estimation Via Patch-Aware Feature Reweighting Under Snowy Conditions, Beibei Zhou, Zhiyuan Zhang, Zhenbo Song, Jianhui Guo, Tongpo Zhang, Shaojing Song, Qinge Wu, Hui Kong

Research Collection School Of Computing and Information Systems

Unsupervised LiDAR motion estimation has attracted significant attention in pattern recognition and robotic vision. However, ensuring robust performance under adverse conditions — particularly structured noise like heavy snowfall — remains a challenging problem. Existing models struggle with domain generalization due to the sensitivity of feature extractors to spatial outliers, limiting their reliability in complex real-world scenarios. In this work, we present an unsupervised LiDAR motion estimation framework via patch-aware feature reweighting to mitigate the domain shift between clear and snowy conditions. Our approach prioritizes reliability-aware reweighting to reduce the influence of snowflake noise and outlier points on pose estimation, while …

The Effects Of Eps Level And Presentation Format Of Analysts’ Forecast Deviation On Non-Professional Investors’ Investment Judgments, Clarence Goh, Prasart Jongjaroenkamol, Poh-Sun Seow

Research Collection School Of Accountancy

We experimentally investigate how the presentation format of the extent to which a firm's earnings per share (EPS) diverges from analysts' EPS forecasts (i.e. deviation information) and a firm's EPS level affect the investment judgments of non-professional investors (referred to hereafter as “investors”). Our results suggest that investors' investment judgments are more positive when firms with low (high) EPS levels disclose deviation information in percentage (absolute) terms. Furthermore, when the percentage of forecast deviation is held constant, investment judgments are more positive when EPS levels are high versus low if the deviation information is expressed in absolute terms. By contrast, …

Generative Ai Adoption And Solvers' Popularity On Supply-Driven Crowdsourcing Platforms: The Dual Role Of Price Signals, Zimeng Zhu, Carol Hsu, Fiona Fui-Hoon Nah, Na Liu

Research Collection School Of Computing and Information Systems

Purpose – We investigate the effect of solvers’ adoption of Generative AI (GenAI) on their popularity in a supply-driven crowdsourcing platform. We also examine the impact of price signals as well as their heterogeneous impact based on the solvers’ membership duration on the platform. Design/methodology/approach – Our analysis focuses on solvers who adopt GenAI for design-related gigs on the supply-driven crowdsourcing platform. By combining propensity score matching (PSM) with multi-period difference-in-differences (DID), we examine how GenAI adoption impacts solvers’ popularity and how price signals affect this main effect. Findings – Our findings reveal that solvers who adopt GenAI tend to …

How Leaders Build Employee Trust In Artificial Intelligence: Voice Opportunities, Humility, And Trust Transfer, Jack Mcguire, David De Cremer, Devesh Narayanan

Research Collection Lee Kong Chian School Of Business

Artificial intelligence is increasingly central to organizational work, yet employee trust in AI remains fragile. Although prior research has primarily explained trust in AI through technological characteristics such as transparency, reliability, and accuracy, we argue that trust in AI is also shaped by the social context in which employees encounter these systems. Drawing on affect-as-information theory and social information processing theory, we develop and test a model in which leader-provided voice opportunities reduce employees’ negative affect about AI-related work experiences, thereby enhancing perceptions of leader trustworthiness and, in turn, trust in AI. We further propose that this indirect effect depends …

The Downstream Effect Of Gender Bias On Academic Performance And Career Aspirations Amongst Female Students In Stem Via Gender-Professional Identity Integration (G-Pii), Chi-Ying Cheng, Shuna Shiann Khoo, Shih-Fen Cheng, Yeow Leong Lee, Vandana Ramachandra Rao

Research Collection School of Social Sciences

Background Gender imbalance in STEM, characterized by a significant underrepresentation of women, remains a significant challenge. Although gender bias is a well-known contributor to women’s attrition from STEM, the psychological mechanisms linking gender bias to departure are less well understood. Our research investigates early antecedents of attrition and the psychological processes that precede leaving the STEM pathway in tertiary education. Drawing upon identity integration research, we propose that perceived gender bias exerts undermines female STEM students’ academic performance and career aspirations by reducing their Gender-Professional Identity Integration (G-PII), a construct that captures individual differences in the perceived compatibility between a …

From Post-Hoc To Pre-Assessment Evaluation: Llms As Collaborative Partners For Simulating Student Performance, Siaw Ling Lo, Ping Fan Ke, Yi Meng Lau

Research Collection School Of Computing and Information Systems

Ensuring the quality of multiple-choice questions (MCQs) remains a complex challenge in higher education, as traditional post-hoc psychometric analyses often identify flaws only after students have been negatively impacted. While Large Language Models (LLMs) offer potential for simulating student performance to validate items before delivery, they are frequently hindered by a “high-accuracy bias” that prevents larger models from realistically mimicking student errors. This paper proposes the Triadic Pedagogical Fit (TPF) framework, which triangulates Model Scale, Task Complexity via Revised Bloom’s Taxonomy, and Instructional Persona to optimize pre-assessment evaluation. Through two studies, we demonstrate that while Mid-scale (12B) models effectively mirror …

Horizontal Regulatory Barriers In International Trade: Evidence From Electric Plugs, Yuan Mei, Mingzhi (Jimmy) Xu

Research Collection School Of Economics

Domestic policies regulating the horizontal dimensions of product differentiation can impede international trade and have become a major concern in contemporary trade negotiations. We estimate trade frictions associated with incompatible designs of electric plugs, a well-known example of a horizontal regulatory barrier. Our findings indicate that fewer electronic devices are exported to destinations with incompatible electric plugs, and this effect operates through both the extensive margin and the intensive margin of exports. These results are further corroborated by an online survey of Chinese electronics exporters. Moreover, incompatible plugs are associated with lower average quality of exported electronic devices, suggesting that …

The Future Of Nature-Based Recreation In Warming Tropical Cities, Perrine Hamel, Emma E. Ramsay, Shawnda A. Morrison, Su Li Heng, Winston T. L. Chow, Beatrice H. Ho, Pearl Min Sze Tan, Moshe Mandelmilch, Lancy Sim, Jason Kai Wei Lee

Research Collection College of Integrative Studies

Nature-based recreation promotes health in tropical cities but is increasingly threatened by rising heat. This review presents recent evidence on the issue of humid heat stress and outdoor recreation in tropical cities and outlines key adaptation strategies – addressing hazard, exposure, and vulnerability – to enable ‘heat-smart’ nature-based activities. Despite existing solutions, critical research gaps remain, especially in integrating social, physiological, and technological insights to better address humid heat stress in tropical urban environments.

Understanding The Out-Of-Distribution Generalization Of Chain-Of-Thought Reasoning In Llms, Nuojing Liang, Xiaotong Yuan, Pan Zhou

Research Collection School Of Computing and Information Systems

Chain-of-Thought (CoT) reasoning substantially improves Large Language Models (LLMs), yet its out-of-distribution (OOD) generalization mechanism remains theoretically underexplored. We study CoT as compositional OOD generalization: the target task lacks complete reasoning trajectories or direct input-output pairs, while training provides only in-distribution (ID) atomic step data. We show that CoT generalizes by repeatedly applying a shared single-step predictor to compose reusable ID transitions. Theoretically, the OOD task risk is controlled by the sum of ID subtask risks, scaled by a depth-independent hard-decoding coefficient Bℓ b = O(k) that is independent of the reasoning depth T. A Rademachercomplexity analysis further gives an …

Team-On-Team Negotiation: How Internal Structure Shapes What Teams Can Do At The Table, Leyi Holly Shi, Tianyu He, Robert B. Lount Jr., Michael Schaerer, Roderick I. Swaab

Research Collection Lee Kong Chian School Of Business

Team negotiation research presents a paradox: teams sometimes achieve more integrative outcomes than solo negotiators, yet in other cases they escalate competitiveness and harm relational outcomes. Existing theory has not resolved this inconsistency, in part because it treats negotiating teams as if they could change positions as easily as individual negotiators. We argue this assumption overlooks a central constraint of team-on-team negotiation: before any adaptive external exchange can occur, teams must internally authorize revisions to their positions as new information emerges. Drawing on evidence from research on team negotiation, multiparty negotiation, and team design, we identify this internal capacity as …

Defense-To-Attack: Bypassing Weak Defenses Enables Stronger Jailbreaks In Vision-Language Models, Yunhan Zhao, Xiang Zheng, Yige Li, Xingjun Ma

Research Collection School Of Computing and Information Systems

Despite their superb capabilities, Vision-Language Models (VLMs) have been shown to be vulnerable to jailbreak attacks. While recent jailbreaks have achieved notable progress, their effectiveness and efficiency can still be improved. In this work, we reveal an interesting phenomenon: incorporating weak defense cues into the attack pipeline can significantly enhance both the effectiveness and efficiency of jailbreaks on VLMs. Building on this insight, we propose Defense2Attack, a novel jailbreak method that bypasses the safety guardrails of VLMs by leveraging defensive patterns to guide jailbreak prompt construction. Specifically, Defense2Attack consists of three key components: (1) a visual optimizer that embeds universal …

Analyzing Developer Discussions On Eu And Us Privacy Legislation Compliance In Github Repositories, Georgia M. Kapitsaki, Maria Papoutsoglou, Christoph Treude, Ioanna Theophilou

Research Collection School Of Computing and Information Systems

Context: Privacy legislation has impacted the way software systems are developed, prompting practitioners to update their implementations. Specifically, the EU General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA) have forced the community to focus on users’ data privacy. Objectives: Relying on the vast amount of data on developer issues available in GitHub repositories, our aim is to gather empirical evidence on the issues developers of Open Source Software discuss to comply with privacy legislation. Method: We examined such discussions by mining and analyzing 32,820 issues from GitHub repositories. We partially analyzed the dataset automatically to identify …

Spatialimaginer: Towards Adaptive Visual Imagination For Spatial Reasoning, Yian Li, Yang Jiao, Bin Zhu, Tianwen Qian, Shaoxiang Chen, Jingjing Chen, Yu-Gang Jiang

Research Collection School Of Computing and Information Systems

Spatial intelligence, which refers to the ability to reason about geometric and physical structure from visual observations, remains a core challenge for multimodal large language models. Despite promising performance, recent multimodal large language models (MLLMs) often exhibit fragile reasoning traces in spatial intelligence tasks that involve consistent spatial state recognition. We argue that these failures stem from a mismatch between the spatial recognition mechanism and the text-only reasoning behavior of these MLLMs. Effective spatial reasoning requires low-level geometric structure to be faithfully preserved and updated throughout the reasoning process, whereas textual representations tend to abstract away precisely these critical details. …

Marching To The Beat: The Role Of Complementor Alignment In The Architectural Evolution Of Ecosystems, Pengxiang Zhang, Liang Chen, Yang Yang, Sali Li

Research Collection Lee Kong Chian School Of Business

The evolution of innovation ecosystems hinges on the effective coordination between ecosystem leaders and complementors. While ecosystem-sponsored architectural innovations create new technological opportunities for complementors, they may unintentionally disrupt the performance of complementor products that were originally developed based on legacy architectural designs. This study provides empirical evidence of these disruptive effects and explores how complementors’ alignment with the ecosystem leader mitigates such challenges. Focusing on Apple’s release of Core ML—a proprietary artificial intelligence module—in its mobile ecosystem, we employ a difference-in-differences design to investigate variations in complementor performance after adopting this ecosystem-sponsored architectural innovation. Our findings reveal that complementors’ …

Climate Transition Stress Testing For Corporate Credit Risk: A Replicable Data-Mining Pipeline Using Esg Risk Scores, Huynh Long Hung Nguyen, Alan @ Ali Madjelisi Megargel, Rajesh Krishna Balan, Paul Robert Griffin

Research Collection School Of Computing and Information Systems

We present a replicable data-mining pipeline for illustrative, scenario-conditioned corporate distress analysis under feature decoupling. The pipeline links Compustat financial statements with Sustainalytics ESG Risk Rating records for 2,144 U.S. non-financial public firms, constructs a standardized Transition Risk Index (TRI), estimates a contemporaneous accounting-distress classifier, and applies transparent transition-cost sensitivities. Distress is defined as annual operating income insufficient to cover annual interest expense; interest coverage itself is excluded from the predictors. The resulting model is not a forward default forecast or a regulatory probability-of-default model. On a held-out cross-sectional test set, the logistic specification obtains AUC 0.909, PR-AUC 0.822, and …

Listening, Partnering, Designing: Smu Libraries’ Journey Towards Human-Centered Wellness, Vincent Ong, Devika Sangaram, Zhenyan Li

Research Collection Library

As student wellness emerges as one of the most pressing concerns facing academic institutions today, libraries are uniquely and powerfully positioned to reimagine their spaces and services as holistic, human-centered environments that nurture the whole person. This paper presents three initiatives by the Singapore Management University (SMU) Libraries that embody this vision, collectively affirming the library's evolving role as a human-centered wellness hub and its commitment to actively shaping a more inclusive and people-first future.

The first initiative repurposed a vacated collection space into a wellbeing-centered quiet study environment, incorporating user experience (UX) design principles and longitudinal survey evidence, to …

User-First Approach To Designing A Library Ai Chatbot, Samantha Seah, Zhenyan Li

Research Collection Library

At SMU Libraries, after a year of experimentation, user testing, and preparation, we have just launched an AI-assisted chatbot to handle Tier 1 questions – will the library be open this weekend, where can I find printers, what workshops does the library offer, etc. While this can be thought of as a technology-forward project, we hope to illustrate how the human elements were essential to elevate the project. Throughout this journey, while learning the ins and outs of new technologies, we placed more emphasis on user-centred service design and we will show how other libraries can do so as well. …

Demonstrating Library Value Through Integrated Research And Data Support In Doctoral Education, Redzuan Abdullah

Research Collection Library

No abstract provided.

Prevalence And Patterns Of Social Isolation And Loneliness Among Older Adults In Singapore, Yi Wen Tan, Le Ha Linh Phan, Emi Loh, Rachel Wen Yi Ngu

ROSA Research Briefs

As Singapore rapidly transitions into a super-aged society (Lau, 2024), social isolation and loneliness have become growing concerns among academics, policymakers, and social service professionals alike (A. Chan et al., 2023; Fang & Tan, 2023; Shafeeq, 2025). News reports in 2025 highlight an increase in helpline calls from older adults seeking help for their loneliness and emotional distress (Shafeeq, 2025). The prevalence of socially isolated elderly living alone is projected to increase fourfold between 2020 and 2060 (Ng et al., 2020), highlighting the urgency and relevance of addressing this social issue.

Unlocking The Potential Of A Superaged Singapore, Yi Wen Tan, Emi Loh, Wensi Lim, Kaung Htet Tun Salai

ROSA Research Briefs

As Singapore enters a new era of longevity, our focus must shift from preparing for longer lives to unlocking their full potential as a societal advantage. This means going beyond adding years to life to ensuring that those years are lived in good health, with meaningful connections and a sense of purpose. Building a longevity-ready society will require communities and institutions alike to adapt, innovate, and create environments where older adults can continue to contribute and participate fully in society. Through research and collaboration, ROSA is committed to advancing knowledge and developing solutions that help Singapore realise the opportunities of …

Work Motivations And Upskilling Preference Among Singapore's Older Adults, Le Ha Linh Phan, Wensi Lim, Emi Loh, Cassandra Ng, Shania Go

ROSA Research Briefs

Singapore's rapidly ageing population and rising life expectancy are reshaping traditional ideas about work, retirement, and learning across the life course. By 2030, around one in four Singapore citizens will be aged 65 and above, reflecting the country's changing demographic structure and increasing longevity (National Population and Talent Division, 2025). As working lives extend, older adults are increasingly navigating a transitional stage between peak career years and later-life retirement, often continuing to remain economically active while simultaneously managing competing priorities such as financial responsibilities, caregiving, health, personal fulfilment, and retirement planning.

Mg-Fuzz: Model-Guided Fuzzing For Unsafe Scenario Discovery In Autonomous Driving Systems, Yulong Lyu, Ruiqi Hong, Jiawan Wang, Jun Sun, Lei Bu

Research Collection School Of Computing and Information Systems

As autonomous driving systems (ADS) are increasingly deployed in real-world environments, discovering diverse unsafe driving scenarios remains a fundamental yet difficult problem. Existing scenario generation and testing approaches often rely on black-box exploration or externally-observed heuristic feedback, which struggle to effectively guide the search toward high-risk scenarios induced by complex decision-making behaviors. A key difficulty stems from the fact that unsafe behaviors in ADS often arise from internal decision-making logic, which can induce structured and discontinuous responses that are hard to effectively explore using purely black-box guidance. Consequently, current tools tend to repeatedly discover a narrow set of similar unsafe …

Chatgpt, Can You Generate Multiple-Choice Questions For An Undergraduate Computing Course? An Evaluation Of Its Effectiveness Using Classical Test Theory, Eng Lieh Ouh, Kar Way Tan, Siaw Ling Lo, Benjamin Gan, Yuhao Zhang

Research Collection School Of Computing and Information Systems

Educators are increasingly adopting large language models such as ChatGPT to generate multiple-choice questions (MCQs), yet empirical evidence on the quality of these items in computing education assessments remains limited. This study compares 30 ChatGPT-generated and 30 manually crafted items from an undergraduate software design course, and evaluates item difficulty, discrimination, and distractor quality using Classical Test Theory. We triangulated two evaluations using within-subjects and between-subjects designs. Our quantitative findings show that all three metrics in both evaluations point in the same direction. ChatGPT-generated items are easier, less discriminating, and exhibit poorer distractor quality than manually crafted items. Distractor quality …

Analyst Visibility And Earnings Forecast Quality, Qiang Cheng, Tian Deng, Sterling Huang, An-Ping Lin

Research Collection School Of Accountancy

This study examines how a reduction in analysts’ visibility, resulting from brokerages’ switch to anonymous forecasts on Eikon, affects earnings forecast quality. Using a difference-in-differences design, we find that treatment analysts respond by issuing more accurate forecasts in the post-anonymization period. Treatment analysts who improve forecast quality the most can sustain visibility, as reflected in their coverage by the financial media, enhance their likelihood of being voted as star analysts, and induce greater client trading. These results are consistent with analysts’ incentives to offset the reduced visibility so as to advance their careers and increase trading commissions for their brokerages. …

Facevalue: Exploring Real-Time Self-View Overlays To Prompt Meaning-Oriented Self-Awareness In Remote Meetings, Gun Woo (Warren) Park, Anthony Tang, Fanny Chevalier

Research Collection School Of Computing and Information Systems

In remote video meetings, visual non-verbal cues, such as facial expressions or head movements, are seen continuously but often only partially. This increases ambiguity compared to in-person settings and can cause misinterpretation or misalignment between intended and perceived meaning. Motivated by communication theories, we designed FaceValue, a technology probe that augments the self-view with private, real-time overlays. These overlays are subtle, suggestive prompts intended to help attendees reflect on how their cues might be interpreted by others. To invite personal interpretation, FaceValue avoids behavioral labeling and instead aims to support meaning-oriented self-awareness: recognizing when visible cues may unintentionally (mis)communicate intent. …

Human Capital Disclosure And Labor Market Outcomes: Evidence From Regulation S-K, Jung Ho Choi, Dan Li, Daniele Macciocchi

Research Collection School Of Accountancy

We examine the labor market consequences of the 2020 Regulation S-K requiring human capital disclosure in 10K filings. Using large-sample job-level data and a Generative Large Language Model (GLLM), we observe that public firms subject to the regulation increase their disclosure of diversity, equity, and inclusion (DEI) information in job postings relative to a matched sample of large private firms. The increase in job-posting disclosure is more pronounced among firms facing greater external pressure to increase their workforce diversity. These findings suggest a shift in demand for diverse candidates by public firms following the regulation. Yet, consistent with short-term inelastic …