Hong Shen

Alessandro Gambetti, Qiwei Han, Cláudia Soares et al.

Vision-Language Models (VLMs) struggle when applied to medical image-text data, yet the tools available to diagnose this failure remain limited. Existing representation alignment metrics are symmetric, collapsing both modalities into a single score and hiding which modality drives cross-modal degradation. We introduce the Spectral Alignment Score (SAS), an asymmetric metric that projects both modalities onto the principal eigenbasis of an anchor modality and computes eigenvalue-weighted per-eigenmode correlations, resulting in directional scores whose difference quantifies modality information imbalance. We embed SAS within a benchmarking framework evaluating 15 VLMs across natural and medical image-text datasets alongside 6 alignment metrics and bidirectional retrieval. Our experiments show that medical images retain richer structural information than their paired clinical reports, a directional asymmetry invisible to all competing metrics, and that SAS achieves the strongest zero-label correlation with retrieval performance in the medical domain, positioning it as a practical diagnostic tool for clinical deployment. Code is available at this URL: https://github.com/iamalegambetti/medical-vlms-assessment.

Wesley Hanwen Deng, Bill Boyuan Guo, Alicia DeVrio et al. · cmu

Recent years have seen growing interest among both researchers and practitioners in user-engaged approaches to algorithm auditing, which directly engage users in detecting problematic behaviors in algorithmic systems. However, we know little about industry practitioners' current practices and challenges around user-engaged auditing, nor what opportunities exist for them to better leverage such approaches in practice. To investigate, we conducted a series of interviews and iterative co-design activities with practitioners who employ user-engaged auditing approaches in their work. Our findings reveal several challenges practitioners face in appropriately recruiting and incentivizing user auditors, scaffolding user audits, and deriving actionable insights from user-engaged audit reports. Furthermore, practitioners shared organizational obstacles to user-engaged auditing, surfacing a complex relationship between practitioners and user auditors. Based on these findings, we discuss opportunities for future HCI research to help realize the potential (and the mitigate risks) of user-engaged auditing in industry practice.

Hong Shen, Ángel Alexander Cabrera, Adam Perer et al. · cmu

This position paper offers a framework to think about how to better involve human influence in algorithmic decision-making of contentious public policy issues. Drawing from insights in communication literature, we introduce a "public(s)-in-the-loop" approach and enumerates three features that are central to this approach: publics as plural political entities, collective decision-making through deliberation, and the construction of publics. It explores how these features might advance our understanding of stakeholder participation in AI design in contentious public policy domains such as recidivism prediction. Finally, it sketches out part of a research agenda for the HCI community to support this work.

Tzu-Sheng Kuo, Hong Shen, Jisoo Geum et al.

Recent years have seen growing adoption of AI-based decision-support systems (ADS) in homeless services, yet we know little about stakeholder desires and concerns surrounding their use. In this work, we aim to understand impacted stakeholders' perspectives on a deployed ADS that prioritizes scarce housing resources. We employed AI lifecycle comicboarding, an adapted version of the comicboarding method, to elicit stakeholder feedback and design ideas across various components of an AI system's design. We elicited feedback from county workers who operate the ADS daily, service providers whose work is directly impacted by the ADS, and unhoused individuals in the region. Our participants shared concerns and design suggestions around the AI system's overall objective, specific model design choices, dataset selection, and use in deployment. Our findings demonstrate that stakeholders, even without AI knowledge, can provide specific and critical feedback on an AI system's design and deployment, if empowered to do so.

Alice Qian, Ryland Shaw, Laura Dabbish et al.

As AI systems are increasingly tested and deployed in open-ended and high-stakes domains, crowdworkers are often tasked with responsible AI (RAI) content work. These tasks include labeling violent content, moderating disturbing text, or simulating harmful behavior for red teaming exercises to shape AI system behaviors. While prior research efforts have highlighted the risks to worker well-being associated with RAI content work, far less attention has been paid to how these risks are communicated to workers by task designers or individuals who design and post RAI tasks. Existing transparency frameworks and guidelines, such as model cards, datasheets, and crowdworksheets, focus on documenting model information and dataset collection processes, but they overlook an important aspect of disclosing well-being risks to workers. In the absence of standard workflows or clear guidance, the consistent application of content warnings, consent flows, or other forms of well-being risk disclosure remains unclear. This study investigates how task designers approach risk disclosure in crowdsourced RAI tasks. Drawing on interviews with 23 task designers across academic and industry sectors, we examine how well-being risk is recognized, interpreted, and communicated in practice. Our findings highlight the need to support task designers in identifying and communicating risks not only to support crowdworker well-being but also to strengthen the ethical integrity and technical efficacy of AI development pipelines.

Wenhao Huang, Haifan Gong, Huan Zhang et al.

CT-based bronchial tree analysis plays an important role in the computer-aided diagnosis for respiratory diseases, as it could provide structured information for clinicians. The basis of airway analysis is bronchial tree reconstruction, which consists of bronchus segmentation and classification. However, there remains a challenge for accurate bronchial analysis due to the individual variations and the severe class imbalance. In this paper, we propose a region and structure prior embedded framework named BronchusNet to achieve accurate segmentation and classification of bronchial regions in CT images. For bronchus segmentation, we propose an adaptive hard region-aware UNet that incorporates multi-level prior guidance of hard pixel-wise samples in the general Unet segmentation network to achieve better hierarchical feature learning. For the classification of bronchial branches, we propose a hybrid point-voxel graph learning module to fully exploit bronchial structure priors and to support simultaneous feature interactions across different branches. To facilitate the study of bronchial analysis, we contribute~\textbf{BRSC}: an open-access benchmark of \textbf{BR}onchus imaging analysis with high-quality pixel-wise \textbf{S}egmentation masks and the \textbf{C}lass of bronchial segments. Experimental results on BRSC show that our proposed method not only achieves the state-of-the-art performance for binary segmentation of bronchial region but also exceeds the best existing method on bronchial branches classification by 6.9\%.

Zijie Fang, Yihan Liu, Yifeng Wang et al.

Biomarker detection is an indispensable part in the diagnosis and treatment of low-grade glioma (LGG). However, current LGG biomarker detection methods rely on expensive and complex molecular genetic testing, for which professionals are required to analyze the results, and intra-rater variability is often reported. To overcome these challenges, we propose an interpretable deep learning pipeline, a Multi-Biomarker Histomorphology Discoverer (Multi-Beholder) model based on the multiple instance learning (MIL) framework, to predict the status of five biomarkers in LGG using only hematoxylin and eosin-stained whole slide images and slide-level biomarker status labels. Specifically, by incorporating the one-class classification into the MIL framework, accurate instance pseudo-labeling is realized for instance-level supervision, which greatly complements the slide-level labels and improves the biomarker prediction performance. Multi-Beholder demonstrates superior prediction performance and generalizability for five LGG biomarkers (AUROC=0.6469-0.9735) in two cohorts (n=607) with diverse races and scanning protocols. Moreover, the excellent interpretability of Multi-Beholder allows for discovering the quantitative and qualitative correlations between biomarker status and histomorphology characteristics. Our pipeline not only provides a novel approach for biomarker prediction, enhancing the applicability of molecular treatments for LGG patients but also facilitates the discovery of new mechanisms in molecular functionality and LGG progression.

Ningjing Tang, Megan Li, Amy Winecoff et al.

Model documentation plays a crucial role in promoting transparency and responsible development of AI systems. With the rise of Generative AI (GenAI), open-source platforms have increasingly become hubs for hosting and distributing these models, prompting platforms like Hugging Face to develop dedicated model documentation guidelines that align with responsible AI principles. Despite these growing efforts, there remains a lack of understanding of how developers document their GenAI models on open-source platforms. Through interviews with 13 GenAI developers active on open-source platforms, we provide empirical insights into their documentation practices and challenges. Our analysis reveals that despite existing resources, developers of GenAI models still face multiple layers of uncertainties in their model documentation: (1) uncertainties about what specific content should be included; (2) uncertainties about how to effectively report key components of their models; and (3) uncertainties in deciding who should take responsibilities for various aspects of model documentation. Based on our findings, we discuss the implications for policymakers, open-source platforms, and the research community to support meaningful, effective and actionable model documentation in the GenAI era, including cultivating better community norms, building robust evaluation infrastructures, and clarifying roles and responsibilities.

Ningjing Tang, Alice Qian, Qiaosi Wang et al.

Content Warning: This paper contains participant quotes and discussions related to mental health challenges, emotional distress, and suicidal ideation. Large language models (LLMs) are increasingly used for mental health support, yet the model safeguards -- particularly refusals to engage with sensitive content -- remain poorly understood from the perspectives of users and mental health professionals (MHPs) and have been reported to cause real-world harms. This paper presents findings from a sequential mixed-methods study examining how LLM refusals are experienced and interpreted in mental health support interactions. Through surveys (N=53) and in-depth interviews (N=16) with individuals using LLMs for mental health support and MHPs, we reveal that refusals are not isolated, single-turn system behaviors but rather constitute dynamic, multi-phase experiences: pre-refusal expectation formation, refusal triggering and encounter, refusal message framing, resource referral provision, and post-refusal outcomes. We contribute a multi-phase framework for evaluating refusals beyond binary policy compliance accuracy and design recommendations for future refusal mechanisms. These findings suggest that understanding LLM refusals requires moving beyond single-turn interactions toward recognizing them as holistic experiences embedded within users' support-seeking trajectories and the broader LLM design pipeline.

Eunkyu Park, Wesley Hanwen Deng, Cheyon Jin et al.

Vision-Language Models (VLMs) continue to struggle to make morally salient judgments in multimodal and socially ambiguous contexts. Prior works typically rely on binary or pairwise supervision, which often fail to capture the continuous and pluralistic nature of human moral reasoning. We present MM-SCALE (Multimodal Moral Scale), a large-scale dataset for aligning VLMs with human moral preferences through 5-point scalar ratings and explicit modality grounding. Each image-scenario pair is annotated with moral acceptability scores and grounded reasoning labels by humans using an interface we tailored for data collection, enabling listwise preference optimization over ranked scenario sets. By moving from discrete to scalar supervision, our framework provides richer alignment signals and finer calibration of multimodal moral reasoning. Experiments show that VLMs fine-tuned on MM-SCALE achieve higher ranking fidelity and more stable safety calibration than those trained with binary signals.

Megan Li, Wendy Bickersteth, Ningjing Tang et al.

Due to its general-purpose nature, Generative AI is applied in an ever-growing set of domains and tasks, leading to an expanding set of risks of harm impacting people, communities, society, and the environment. These risks may arise due to failures during the design and development of the technology, as well as during its release, deployment, or downstream usages and appropriations of its outputs. In this paper, building on prior taxonomies of AI risks, harms, and failures, we construct a taxonomy specifically for Generative AI failures and map them to the harms they precipitate. Through a systematic analysis of 499 publicly reported incidents, we describe what harms are reported, how they arose, and who they impact. We report the prevalence of each type of harm, underlying failure mode, and harmed stakeholder, as well as their common co-occurrences. We find that most reported incidents are caused by use-related issues but bring harm to parties beyond the end user(s) of the Generative AI system at fault, and that the landscape of Generative AI harms is distinct from that of traditional AI. Our work offers actionable insights to policymakers, developers, and Generative AI users. In particular, we call for the prioritization of non-technical risk and harm mitigation strategies, including public disclosures and education and careful regulatory stances.

Rongyi Chen, Ziyan Xin, Qing Xiao et al.

The digital transformation of religious practice has reshaped how billions of people engage with spiritual content, with video-sharing platforms becoming central to contemporary religious communication. Yet HCI research lacks systematic understanding of how narrative and visual elements create meaningful spiritual experiences and foster viewer engagement. We present a mixed-methods study of religious videos on YouTube across major religions, developing taxonomies of narrative frameworks, visual elements, and viewer interaction. Using LLM-assisted analysis, we studied relationships between content characteristics and viewer responses. Religious videos predominantly adopt lecture-style formats with authority-based persuasion strategies, using salvation narratives for guidance. All prefer bright lighting, with Buddhism favoring warm tones and prominent symbols, Judaism preferring indoor settings, and Hinduism emphasizing sacred objects. We identified differentiated patterns of emotional sharing among religious viewers while revealing significant correlations between content characteristics and engagement, particularly regarding AI-generated content. We provide evidence-based guidance for creating inclusive and engaging spiritual media.

Haifan Gong, Wenhao Huang, Huan Zhang et al.

Automatic segmentation of the bronchial tree from CT imaging is important, as it provides structural information for disease diagnosis. Despite the merits of previous automatic bronchus segmentation methods, they have paied less attention to the issue we term as \textit{Intensity Confusion}, wherein the intensity values of certain background voxels approach those of the foreground voxels within bronchi. Conversely, the intensity values of some foreground voxels are nearly identical to those of background voxels. This proximity in intensity values introduces significant challenges to neural network methodologies. To address the issue, we introduce a novel Intensity-Distance Guided loss function, which assigns adaptive weights to different image voxels for mining hard samples that cause the intensity confusion. The proposed loss estimates the voxel-level hardness of samples, on the basis of the following intensity and distance priors. We regard a voxel as a hard sample if it is in: (1) the background and has an intensity value close to the bronchus region; (2) the bronchus region and is of higher intensity than most voxels inside the bronchus; (3) the background region and at a short distance from the bronchus. Extensive experiments not only show the superiority of our method compared with the state-of-the-art methods, but also verify that tackling the intensity confusion issue helps to significantly improve bronchus segmentation. Project page: https://github.com/lhaof/ICM.

Yee-Yin Choong, Kristen Greene, Alice Qian et al.

AI measurement science has a wide variety of methodologies and measurements for comparing AI systems, resulting in what often appear to be "apples-to-oranges" comparisons across AI evaluations. To move toward "apples-to-apples" comparisons in real-world AI evaluations, this work advocates for methodological transparency in evaluation scenarios, operational grounding, and human-centered design (HCD) principles. We propose a repeatable process for transforming high-level use cases to detailed scenarios by eliciting use cases from subject matter experts (SMEs) via a structured AI Use Case Worksheet with six key elements: use case, sector, user (direct and indirect), intended outcomes, expected impacts (positive and negative), and KPIs and metrics. We demonstrate utility of the worksheet and process in the U.S. financial services sector. This paper reports on example high-level AI use cases identified by financial services sector SMEs: cyber defense enablement, developer productivity, financial crime aggregation, suspicious activity report (SAR) filing, credit memo generation, and internal call center support. These AI use cases provided are illustrative of the process and not exhaustive. Central to our work is a three-stage expansion pipeline combining LLM prompting with human reviews to generate 107 scenarios from those use cases elicited from SMEs. This process integrates iterative human reviews at every juncture to ensure operational grounding: for scenario titles and descriptions; for core scenario elements like users, benefits and risks, and metrics; and for scenario narratives and evaluation objectives. Human checkpoints ensure scenarios remain reflective of real-world usage and human needs. We describe a validation rubric to assess scenario quality. By defining key scenario components, this work supports a more consistent and meaningful paradigm for human-centered AI evaluations.

Yike Shi, Qing Xiao, Qing Hu et al.

Large language models can influence users through conversation, creating new forms of dark patterns that differ from traditional UX dark patterns. We define LLM dark patterns as manipulative or deceptive behaviors enacted in dialogue. Drawing on prior work and AI incident reports, we outline a diverse set of categories with real-world examples. Using them, we conducted a scenario-based study where participants (N=34) compared manipulative and neutral LLM responses. Our results reveal that recognition of LLM dark patterns often hinged on conversational cues such as exaggerated agreement, biased framing, or privacy intrusions, but these behaviors were also sometimes normalized as ordinary assistance. Users' perceptions of these dark patterns shaped how they respond to them. Responsibilities for these behaviors were also attributed in different ways, with participants assigning it to companies and developers, the model itself, or to users. We conclude with implications for design, advocacy, and governance to safeguard user autonomy.

Cindy Peng, Megan Chai, Gao Mo et al.

Peer-run organizations (PROs) provide critical, recovery-based behavioral health support rooted in lived experience. As large language models (LLMs) enter this domain, their scale, conversationality, and opacity introduce new challenges for situatedness, trust, and autonomy. Partnering with Collaborative Support Programs of New Jersey (CSPNJ), a statewide PRO in the Northeastern United States, we used comicboarding, a co-design method, to conduct workshops with 16 peer specialists and 10 service users exploring perceptions of integrating an LLM-based recommendation system into peer support. Findings show that depending on how LLMs are introduced, constrained, and co-used, they can reconfigure in-room dynamics by sustaining, undermining, or amplifying the relational authority that grounds peer support. We identify opportunities, risks, and mitigation strategies across three tensions: bridging scale and locality, protecting trust and relational dynamics, and preserving peer autonomy amid efficiency gains. We contribute design implications that center lived-experience-in-the-loop, reframe trust as co-constructed, and position LLMs not as clinical tools but as relational collaborators in high-stakes, community-led care.

Megan Li, Wendy Bickersteth, Ningjing Tang et al.

Despite growing concerns about the risks of Generative AI (GenAI), there is limited understanding of public perceptions of these risks and their associated failure modes -- defined as recurring patterns of sociotechnical breakdown across the GenAI lifecycle that contribute to risks of real-world harm. To address this gap, we present a survey instrument, validated with eight subject matter experts and deployed on a sample of 960 U.S.-based participants, to assess awareness and perceptions of GenAI's failure modes, their associated risks, and stakeholder responsibilities to address them. To support realism and content validity, our instrument is structured around scenarios grounded in publicly reported incidents and a taxonomy of GenAI's failure modes. Findings suggest that our instrument is (1) effective for assessing risk awareness and perceptions in a way that is grounded in people's current contexts of use, yet is extensible to new contexts that will inevitably arise; and (2) potentially useful for informing the design of AI literacy tools and interventions. We argue for AI literacy and governance approaches that align with how people encounter and reason about GenAI in everyday life.

Xin Wang, Hong Shen, Hui Tian et al.

Coflow has emerged as a fundamental application-layer abstraction in distributed systems, representing communication dependencies and enabling collaborative management of related flows to enhance job completion efficiency. To meet the increasing bandwidth demands of modern data center networks (DCNs), optical circuit switches are widely deployed due to their high capacity and energy efficiency. Simultaneously, DCN deployments are evolving towards heterogeneous parallel architectures, where multiple independent optical circuit switching (OCS) cores operate concurrently to facilitate bandwidth expansion and incremental upgrades. However, existing research on coflow scheduling in multi-core switching fabrics primarily focuses on electrical packet switching (EPS) networks, with a few known results on OCS networks without or with a poor performance guarantee. This paper studies the coflow scheduling problem in multi-core OCS networks under the not-all-stop (i.e., asynchronous) reconfiguration model, focusing on two major challenges of overcoming cross-core coupling for inter-core traffic allocation and satisfying the constraints of port exclusivity and reconfiguration overhead for intra-core circuit scheduling. To minimize total weighted coflow completion time (CCT), we propose an efficient algorithm by integrating linear programming-guided (LP-guided) global coflow ordering, inter-core flow allocation and intra-core circuit scheduling that achieves approximation ratios of 8K and 8K+1 for zero and arbitrary release times of coflows, respectively, where K is the number of OCS cores. This framework is also applicable to H-core EPS networks, providing approximation guarantees of 4H and 4H+1 for zero-time and arbitrary-time release, respectively.

Qiaosi Wang, Xuhui Zhou, Maarten Sap et al. · allen-ai, cmu

The last couple of years have witnessed emerging research that appropriates Theory-of-Mind (ToM) tasks designed for humans to benchmark LLM's ToM capabilities as an indication of LLM's social intelligence. However, this approach has a number of limitations. Drawing on existing psychology and AI literature, we summarize the theoretical, methodological, and evaluation limitations by pointing out that certain issues are inherently present in the original ToM tasks used to evaluate human's ToM, which continues to persist and exacerbated when appropriated to benchmark LLM's ToM. Taking a human-computer interaction (HCI) perspective, these limitations prompt us to rethink the definition and criteria of ToM in ToM benchmarks in a more dynamic, interactional approach that accounts for user preferences, needs, and experiences with LLMs in such evaluations. We conclude by outlining potential opportunities and challenges towards this direction.

Jingwen Cheng, Kshitish Ghate, Wenyue Hua et al.

Large Language Models (LLMs), such as the GPT series, have driven significant industrial applications, leading to economic and societal transformations. However, a comprehensive understanding of their real-world applications remains limited. To address this, we introduce REALM, a dataset of over 94,000 LLM use cases collected from Reddit and news articles. REALM captures two key dimensions: the diverse applications of LLMs and the demographics of their users. It categorizes LLM applications and explores how users' occupations relate to the types of applications they use. By integrating real-world data, REALM offers insights into LLM adoption across different domains, providing a foundation for future research on their evolving societal roles.

Alessandro Gambetti, Qiwei Han, Hong Shen et al.

Explainable Artificial Intelligence (XAI) is essential for the transparency and clinical adoption of Clinical Decision Support Systems (CDSS). However, the real-world effectiveness of existing XAI methods remains limited and is inconsistently evaluated. This study conducts a systematic PRISMA-guided survey of 31 human-centered evaluations (HCE) of XAI applied to CDSS, classifying them by XAI methodology, evaluation design, and adoption barrier. Our findings reveal that most existing studies employ post-hoc, model-agnostic approaches such as SHAP and Grad-CAM, typically assessed through small-scale clinician studies. The results show that over 80% of the studies adopt post-hoc, model-agnostic approaches such as SHAP and Grad-CAM, and that clinician sample sizes remain below 25 participants. The findings indicate that explanations generally improve clinician trust and diagnostic confidence, but frequently increase cognitive load and exhibit misalignment with domain reasoning processes. To bridge these gaps, we propose a stakeholder-centric evaluation framework that integrates socio-technical principles and human-computer interaction to guide the future development of clinically viable and trustworthy XAI-based CDSS.

Jian Wang, Hong Shen, Chan-Tong Lam

Traditional distributed backdoor attacks (DBA) in federated learning improve stealthiness by decomposing global triggers into sub-triggers, which however requires more poisoned data to maintian the attck strength and hence increases the exposure risk. To overcome this defect, This paper proposes a novel method, namely Fractal-Triggerred Distributed Backdoor Attack (FTDBA), which leverages the self-similarity of fractals to enhance the feature strength of sub-triggers and hence significantly reduce the required poisoning volume for the same attack strength. To address the detectability of fractal structures in the frequency and gradient domains, we introduce a dynamic angular perturbation mechanism that adaptively adjusts perturbation intensity across the training phases to balance efficiency and stealthiness. Experiments show that FTDBA achieves a 92.3\% attack success rate with only 62.4\% of the poisoning volume required by traditional DBA methods, while reducing the detection rate by 22.8\% and KL divergence by 41.2\%. This study presents a low-exposure, high-efficiency paradigm for federated backdoor attacks and expands the application of fractal features in adversarial sample generation.

Xin Wang, Hong Shen, Hui Tian et al.

Coflow provides a key application-layer abstraction for capturing communication patterns, enabling the efficient coordination of parallel data flows to reduce job completion times in distributed systems. Modern data center networks (DCNs) are employing multiple independent optical circuit switching (OCS) cores operating concurrently to meet the massive bandwidth demands of application jobs. However, existing coflow scheduling research primarily focuses on the single-core setting, with multi-core fabrics only for EPS (electrical packet switching) networks. To address this gap, this paper studies the coflow scheduling problem in multi-core OCS networks under the not-all-stop reconfiguration model in which one circuit's reconfiguration does not interrupt other circuits. The challenges stem from two aspects: (i) cross-core coupling induced by traffic assignment across heterogeneous cores; and (ii) per-core OCS scheduling constraints, namely port exclusivity and reconfiguration delay. We propose an approximation algorithm that jointly integrates cross-core flow assignment and per-core circuit scheduling to minimize the total weighted coflow completion time (CCT) and establish a provable worst-case performance guarantee. Furthermore, our algorithm framework can be directly applied to the multi-core EPS scenario with the corresponding approximation ratio under packet-switched fabrics. Trace-driven simulations using real Facebook workloads demonstrate that our algorithm effectively reduces weighted CCT and tail CCT.

Muzhe Wu, Yanzhi Zhao, Shuyi Han et al.

Responsible AI (RAI) efforts increasingly emphasize the importance of addressing potential harms early in the AI development lifecycle through social-technical lenses. However, in cross-functional industry teams, this work is often stalled by a persistent knowledge handoff challenge: the difficulty of transferring high-level, early-stage technical design rationales from technical experts to non-technical or user-facing roles for ethical evaluation and harm identification. Through literature review and a co-design study with 8 practitioners, we unpack how this challenge manifests -- technical design choices are rarely handed off in ways that support meaningful engagement by non-technical roles; collaborative workflows lack shared, visual structures to support mutual understanding; and non-technical practitioners are left without scaffolds for systematic harm evaluation. Existing tools like JIRA or Google Docs, while useful for product tracking, are ill-suited for supporting joint harm identification across roles, often requiring significant extra effort to align understanding. To address this, we developed AI LEGO, a web-based prototype that supports cross-functional AI practitioners in effectively facilitating knowledge handoff and identifying harmful design choices in the early design stages. Technical roles use interactive blocks to draft development plans, while non-technical roles engage with those blocks through stage-specific checklists and LLM-driven persona simulations to surface potential harms. In a study with 18 cross-functional practitioners, AI LEGO increased the volume and likelihood of harms identified compared to baseline worksheets. Participants found that its modular structure and persona prompts made harm identification more accessible, fostering clearer and more collaborative RAI practices in early design.

Evans Xu Han, Alice Qian Zhang, Haiyi Zhu et al.

State-of-the-art visual generative AI tools hold immense potential to assist users in the early ideation stages of creative tasks -- offering the ability to generate (rather than search for) novel and unprecedented (instead of existing) images of considerable quality that also adhere to boundless combinations of user specifications. However, many large-scale text-to-image systems are designed for broad applicability, yielding conventional output that may limit creative exploration. They also employ interaction methods that may be difficult for beginners. Given that creative end users often operate in diverse, context-specific ways that are often unpredictable, more variation and personalization are necessary. We introduce POET, a real-time interactive tool that (1) automatically discovers dimensions of homogeneity in text-to-image generative models, (2) expands these dimensions to diversify the output space of generated images, and (3) learns from user feedback to personalize expansions. An evaluation with 28 users spanning four creative task domains demonstrated POET's ability to generate results with higher perceived diversity and help users reach satisfaction in fewer prompts during creative tasks, thereby prompting them to deliberate and reflect more on a wider range of possible produced results during the co-creative process. Focusing on visual creativity, POET offers a first glimpse of how interaction techniques of future text-to-image generation tools may support and align with more pluralistic values and the needs of end users during the ideation stages of their work.

Xianzhe Fan, Qing Xiao, Xuhui Zhou et al. · allen-ai, cmu

AI companions based on large language models can role-play and converse very naturally. When value conflicts arise between the AI companion and the user, it may offend or upset the user. Yet, little research has examined such conflicts. We first conducted a formative study that analyzed 151 user complaints about conflicts with AI companions, providing design implications for our study. Based on these, we created Minion, a technology probe to help users resolve human-AI value conflicts. Minion applies a user-empowerment intervention method that provides suggestions by combining expert-driven and user-driven conflict resolution strategies. We conducted a technology probe study, creating 40 value conflict scenarios on Character.AI and Talkie. 22 participants completed 274 tasks and successfully resolved conflicts 94.16% of the time. We summarize user responses, preferences, and needs in resolving value conflicts, and propose design implications to reduce conflicts and empower users to resolve them more effectively.

Alice Qian, Ziqi Yang, Ryland Shaw et al.

Responsible AI (RAI) content work, such as annotation, moderation, or red teaming for AI safety, often exposes crowd workers to potentially harmful content. While prior work has underscored the importance of communicating well-being risk to employed content moderators, designing effective disclosure mechanisms for crowd workers while balancing worker protection with the needs of task designers and platforms remains largely unexamined. To address this gap, we conducted individual co-design sessions with 15 task designers, 11 crowdworkers, and 3 platform representatives. We investigated task designer preferences for support in disclosing tasks, worker preferences for receiving risk disclosure warnings, and how platform representatives envision their role in shaping risk disclosure practices. We identify design tensions and map the sociotechnical tradeoffs that shape disclosure practices. We contribute design recommendations and feature concepts for risk disclosure mechanisms in the context of RAI content work.

Qiaosi Wang, Jini Kim, Avanita Sharma et al.

Theory of Mind (ToM) -- the ability to infer what others are thinking (e.g., intentions) from observable cues -- is traditionally considered fundamental to human social interactions. This has sparked growing efforts in building and benchmarking AI's ToM capability, yet little is known about how such capability could translate into the design and experience of everyday user-facing AI products and services. We conducted 13 co-design sessions with 26 U.S.-based AI practitioners to envision, reflect, and distill design recommendations for ToM-enabled everyday AI products and services that are both future-looking and grounded in the realities of AI design and development practices. Analysis revealed three interrelated design recommendations: ToM-enabled AI should 1) be situated in the social context that shape users' mental states, 2) be responsive to the dynamic nature of mental states, and 3) be attuned to subjective individual differences. We surface design tensions within each recommendation that reveal a broader gap between practitioners' envisioned futures of ToM-enabled AI and the realities of current AI design and development practices. These findings point toward the need to move beyond static, inference-driven approach to ToM and toward designing ToM as a pervasive capability that supports continuous human-AI interaction loops.

Ruiwei Xiao, Qing Xiao, Xinying Hou et al.

Generative AI (GenAI) is rapidly entering K-12 classrooms, offering teachers new ways for teaching practices. Yet GenAI models are often trained on culturally uneven datasets, embedding a "default culture" that often misaligns with local classrooms. To understand how teachers navigate this gap, we defined the new concept Cultural Distance (the gap between GenAI's default cultural repertoire and the situated demands of teaching practice) and conducted in-depth interviews with 30 K-12 teachers, 10 each from South Africa, Taiwan, and the United States, who had integrated AI into their teaching practice. These teachers' experiences informed the development of our three-level cultural distance framework. This work contributes the concept and framework of cultural distance, six illustrative instances spanning in low, mid, high distance levels with teachers' experiences and strategies for addressing them. Empirically, we offer implications to help AI designers, policymakers, and educators create more equitable and culturally responsive GenAI tools for education.

Fangzhou Dong, Yifan Zeng, Yingpeng Sang et al.

Large Language Models (LLMs) excel in understanding and generating text but struggle with providing professional literary criticism for works with profound thoughts and complex narratives. This paper proposes GLASS (Greimas Literary Analysis via Semiotic Square), a structured analytical framework based on Greimas Semiotic Square (GSS), to enhance LLMs' ability to conduct in-depth literary analysis. GLASS facilitates the rapid dissection of narrative structures and deep meanings in narrative works. We propose the first dataset for GSS-based literary criticism, featuring detailed analyses of 48 works. Then we propose quantitative metrics for GSS-based literary criticism using the LLM-as-a-judge paradigm. Our framework's results, compared with expert criticism across multiple works and LLMs, show high performance. Finally, we applied GLASS to 39 classic works, producing original and high-quality analyses that address existing research gaps. This research provides an AI-based tool for literary research and education, offering insights into the cognitive mechanisms underlying literary engagement.

Xin Wang, Hong Shen

Coflow is a recently proposed networking abstraction to help improve the communication performance of data-parallel computing jobs. In multi-stage jobs, each job consists of multiple coflows and is represented by a Directed Acyclic Graph (DAG). Efficiently scheduling coflows is critical to improve the data-parallel computing performance in data centers. Compared with hand-tuned scheduling heuristics, existing work DeepWeave [1] utilizes Reinforcement Learning (RL) framework to generate highly-efficient coflow scheduling policies automatically. It employs a graph neural network (GNN) to encode the job information in a set of embedding vectors, and feeds a flat embedding vector containing the whole job information to the policy network. However, this method has poor scalability as it is unable to cope with jobs represented by DAGs of arbitrary sizes and shapes, which requires a large policy network for processing a high-dimensional embedding vector that is difficult to train. In this paper, we first utilize a directed acyclic graph neural network (DAGNN) to process the input and propose a novel Pipelined-DAGNN, which can effectively speed up the feature extraction process of the DAGNN. Next, we feed the embedding sequence composed of schedulable coflows instead of a flat embedding of all coflows to the policy network, and output a priority sequence, which makes the size of the policy network depend on only the dimension of features instead of the product of dimension and number of nodes in the job's DAG.Furthermore, to improve the accuracy of the priority scheduling policy, we incorporate the Self-Attention Mechanism into a deep RL model to capture the interaction between different parts of the embedding sequence to make the output priority scores relevant. Based on this model, we then develop a coflow scheduling algorithm for online multi-stage jobs.

Hong Shen, Alicia DeVos, Motahhare Eslami et al.

A growing body of literature has proposed formal approaches to audit algorithmic systems for biased and harmful behaviors. While formal auditing approaches have been greatly impactful, they often suffer major blindspots, with critical issues surfacing only in the context of everyday use once systems are deployed. Recent years have seen many cases in which everyday users of algorithmic systems detect and raise awareness about harmful behaviors that they encounter in the course of their everyday interactions with these systems. However, to date little academic attention has been granted to these bottom-up, user-driven auditing processes. In this paper, we propose and explore the concept of everyday algorithm auditing, a process in which users detect, understand, and interrogate problematic machine behaviors via their day-to-day interactions with algorithmic systems. We argue that everyday users are powerful in surfacing problematic machine behaviors that may elude detection via more centrally-organized forms of auditing, regardless of users' knowledge about the underlying algorithms. We analyze several real-world cases of everyday algorithm auditing, drawing lessons from these cases for the design of future platforms and tools that facilitate such auditing behaviors. Finally, we discuss work that lies ahead, toward bridging the gaps between formal auditing approaches and the organic auditing behaviors that emerge in everyday use of algorithmic systems.

Luodi Xie, Hong Shen, Jiaxin Ren

Network embedding maps the nodes of a given network into a low-dimensional space such that the semantic similarities among the nodes can be effectively inferred. Most existing approaches use inner-product of node embedding to measure the similarity between nodes leading to the fact that they lack the capacity to capture complex relationships among nodes. Besides, they take the path in the network just as structural auxiliary information when inferring node embeddings, while paths in the network are formed with rich user informations which are semantically relevant and cannot be ignored. In this paper, We propose a novel method called Network Embedding on the Metric of Relation, abbreviated as NEMR, which can learn the embeddings of nodes in a relational metric space efficiently. First, our NEMR models the relationships among nodes in a metric space with deep learning methods including variational inference that maps the relationship of nodes to a gaussian distribution so as to capture the uncertainties. Secondly, our NEMR considers not only the equivalence of multiple-paths but also the natural order of a single-path when inferring embeddings of nodes, which makes NEMR can capture the multiple relationships among nodes since multiple paths contain rich user information, e.g., age, hobby and profession. Experimental results on several public datasets show that the NEMR outperforms the state-of-the-art methods on relevant inference tasks including link prediction and node classification.

Hong Shen, Wesley Hanwen Deng, Aditi Chattopadhyay et al.

Recently, there have been increasing calls for computer science curricula to complement existing technical training with topics related to Fairness, Accountability, Transparency, and Ethics. In this paper, we present Value Card, an educational toolkit to inform students and practitioners of the social impacts of different machine learning models via deliberation. This paper presents an early use of our approach in a college-level computer science course. Through an in-class activity, we report empirical data for the initial effectiveness of our approach. Our results suggest that the use of the Value Cards toolkit can improve students' understanding of both the technical definitions and trade-offs of performance metrics and apply them in real-world contexts, help them recognize the significance of considering diverse social values in the development of deployment of algorithmic systems, and enable them to communicate, negotiate and synthesize the perspectives of diverse stakeholders. Our study also demonstrates a number of caveats we need to consider when using the different variants of the Value Cards toolkit. Finally, we discuss the challenges as well as future applications of our approach.

Zhigang Lu, Hong Shen

Iterative clustering algorithms help us to learn the insights behind the data. Unfortunately, this may allow adversaries to infer the privacy of individuals with some background knowledge. In the worst case, the adversaries know the centroids of an arbitrary iteration and the information of n-1 out of n items. To protect individual privacy against such an inference attack, preserving differential privacy (DP) for the iterative clustering algorithms has been extensively studied in the interactive settings. However, existing interactive differentially private clustering algorithms suffer from a non-convergence problem, i.e., these algorithms may not terminate without a predefined number of iterations. This problem severely impacts the clustering quality and the efficiency of a differentially private algorithm. To resolve this problem, in this paper, we propose a novel differentially private clustering framework in the interactive settings which controls the orientation of the movement of the centroids over the iterations to ensure the convergence by injecting DP noise in a selected area. We prove that, in the expected case, algorithm under our framework converges in at most twice the iterations of Lloyd's algorithm. We perform experimental evaluations on real-world datasets to show that our algorithm outperforms the state-of-the-art of the interactive differentially private clustering algorithms with guaranteed convergence and better clustering quality to meet the same DP requirement.

Zhigang Lu, Hong Shen

Paths in a given network are a generalised form of time-serial chains in many real-world applications, such as trajectories and Internet flows. Differentially private trajectory publishing concerns publishing path information that is usable to the genuine users yet secure against adversaries to reconstruct the path with maximum background knowledge. The exiting studies all assume this knowledge to be all but one vertex on the path. To prevent the adversaries recovering the missing information, they publish a perturbed path where each vertex is sampled from a pre-defined set with differential privacy (DP) to replace the corresponding vertex in the original path. In this paper, we relax this assumption to be all but one edge on the path, and hence consider the scenario of more powerful adversaries with the maximum background knowledge of the entire network topology and the path (including all the vertices) except one (arbitrary) missing edge. Under such an assumption, the perturbed path produced by the existing work is vulnerable, because the adversary can reconstruct the missing edge from the existence of an edge in the perturbed path. To address this vulnerability and effectively protect edge privacy, instead of publishing a perturbed path, we propose a novel scheme of graph-based path publishing to protect the original path by embedding the path in a graph that contains fake edges and replicated vertices applying the differential privacy technique, such that only the legitimate users who have the full knowledge of the network topology are able to recover the exact vertices and edges of the original path with high probability. We theoretically analyse the performance of our algorithm in differential privacy, utility, and execution efficiency. We also conduct extensive experimental evaluations on a high-performance cluster system to validate our analytical results.

Teng Xiao, Shangsong Liang, Hong Shen et al.

Collaborative Filtering (CF) is one of the most used methods for Recommender System. Because of the Bayesian nature and nonlinearity, deep generative models, e.g. Variational Autoencoder (VAE), have been applied into CF task, and have achieved great performance. However, most VAE-based methods suffer from matrix sparsity and consider the prior of users' latent factors to be the same, which leads to poor latent representations of users and items. Additionally, most existing methods model latent factors of users only and but not items, which makes them not be able to recommend items to a new user. To tackle these problems, we propose a Neural Variational Hybrid Collaborative Filtering, NVHCF. Specifically, we consider both the generative processes of users and items, and the prior of latent factors of users and items to be side informationspecific, which enables our model to alleviate matrix sparsity and learn better latent representations of users and items. For inference purpose, we derived a Stochastic Gradient Variational Bayes (SGVB) algorithm to analytically approximate the intractable distributions of latent factors of users and items. Experiments conducted on two large datasets have showed our methods significantly outperform the state-of-the-art CF methods, including the VAE-based methods.

Zaiqiao Meng, Hong Shen

Given a large network and a query node, finding its top-k similar nodes is a primitive operation in many graph-based applications. Recently enhancing search results with diversification have received much attention. In this paper, we explore an novel problem of searching for top-k diversified similar nodes in attributed networks, with the motivation that modeling diversification in an attributed network should consider both the emergence of network links and the attribute features of nodes such as user profile information. We formulate this practical problem as two optimization problems: the Attributed Coverage Diversification (ACD) problem and the r-Dissimilar Attributed Coverage Diversification (r-DACD) problem. Based on the submodularity and the monotonicity of ACD, we propose an efficient greedy algorithm achieving a tight approximation guarantee of 1-1/e. Unlike the expension based methods only considering nodes' neighborhood, ACD generalize the definition of diversification to nodes' own features. To capture diversification in topological structure of networks, the r-DACD problem introduce a dissimilarity constraint. We refer to this problem as the Dissimilarity Constrained Non-monotone Submodular Maximization (DCNSM) problem. We prove that there is no constant-factor approximation for DCNSM, and also present an efficient greedy algorithms achieving $1/ρ$ approximation, where $ρ\leΔ$, $Δ$ is the maximum degree of its dissimilarity based graph. To the best of our knowledge, it is the first approximation algorithm for the Submodular Maximization problem with a distance constraint. The experimental results on real-world attributed network datasets demonstrate the effectiveness of our methods, and confirm that adding dissimilarity constraint can significantly enhance the performance of diversification.

Zhigang Lu, Hong Shen

Neighbourhood-based Collaborative Filtering (CF) has been applied in the industry for several decades, because of the easy implementation and high recommendation accuracy. As the core of neighbourhood-based CF, the task of dynamically maintaining users' similarity list is challenged by cold-start problem and scalability problem. Recently, several methods are presented on solving the two problems. However, these methods applied an $O(n^2)$ algorithm to compute the similarity list in a special case, where the new users, with enough recommendation data, have the same rating list. To address the problem of large computational cost caused by the special case, we design a faster ($O(\frac{1}{125}n^2)$) algorithm, TwinSearch Algorithm, to avoid computing and sorting the similarity list for the new users repeatedly to save the computational resources. Both theoretical and experimental results show that the TwinSearch Algorithm achieves better running time than the traditional method.

Zhigang Lu, Hong Shen

The neighbourhood-based Collaborative Filtering is a widely used method in recommender systems. However, the risks of revealing customers' privacy during the process of filtering have attracted noticeable public concern recently. Specifically, $k$NN attack discloses the target user's sensitive information by creating $k$ fake nearest neighbours by non-sensitive information. Among the current solutions against $k$NN attack, the probabilistic methods showed a powerful privacy preserving effect. However, the existing probabilistic methods neither guarantee enough prediction accuracy due to the global randomness, nor provide assured security enforcement against $k$NN attack. To overcome the problems of current probabilistic methods, we propose a novel approach, Partitioned Probabilistic Neighbour Selection, to ensure a required security guarantee while achieving the optimal prediction accuracy against $k$NN attack. In this paper, we define the sum of $k$ neighbours' similarity as the accuracy metric $α$, the number of user partitions, across which we select the $k$ neighbours, as the security metric $β$. Differing from the present methods that globally selected neighbours, our method selects neighbours from each group with exponential differential privacy to decrease the magnitude of noise. Theoretical and experimental analysis show that to achieve the same security guarantee against $k$NN attack, our approach ensures the optimal prediction accuracy.

Zhigang Lu, Hong Shen

Recommender systems, tool for predicting users' potential preferences by computing history data and users' interests, show an increasing importance in various Internet applications such as online shopping. As a well-known recommendation method, neighbourhood-based collaborative filtering has attracted considerable attention recently. The risk of revealing users' private information during the process of filtering has attracted noticeable research interests. Among the current solutions, the probabilistic techniques have shown a powerful privacy preserving effect. When facing $k$ Nearest Neighbour attack, all the existing methods provide no data utility guarantee, for the introduction of global randomness. In this paper, to overcome the problem of recommendation accuracy loss, we propose a novel approach, Partitioned Probabilistic Neighbour Selection, to ensure a required prediction accuracy while maintaining high security against $k$NN attack. We define the sum of $k$ neighbours' similarity as the accuracy metric alpha, the number of user partitions, across which we select the $k$ neighbours, as the security metric beta. We generalise the $k$ Nearest Neighbour attack to beta k Nearest Neighbours attack. Differing from the existing approach that selects neighbours across the entire candidate list randomly, our method selects neighbours from each exclusive partition of size $k$ with a decreasing probability. Theoretical and experimental analysis show that to provide an accuracy-assured recommendation, our Partitioned Probabilistic Neighbour Selection method yields a better trade-off between the recommendation accuracy and system security.