Lindy Le

Lindy Le

Auto-generated profile, based on publicly available information

Lindy Le

University of Michigan

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
24 September 2024 in ACM Transactions on Interactive Intelligent Systems
Verified
Journal Article
Open access
CLConnor Lawless
Connor Lawless
JSJakob Schoeffer
Jakob Schoeffer
LLLindy Le
Lindy Le
KRKael Rowan
Kael Rowan
SSShilad Sen
Shilad Sen
BSBahareh Sarrafzadeh
Bahareh Sarrafzadeh
A critical factor in the success of many decision support systems is the accurate modeling of user preferences. Psychology research has demonstrated that users often develop their preferences during the elicitation process, highlighting the pivotal role of system-user interaction in developing personalized systems. This paper introduces a novel approach, combining Large Language Models (LLMs) with Constraint Programming to facilitate interactive decision support. We study this hybrid framework through the lens of meeting scheduling, a time-consuming daily activity faced by a multitude of information workers. We conduct three studies to evaluate the novel framework, including a diary study to characterize contextual scheduling preferences, a quantitative evaluation of the system’s performance, and a user study to elicit insights with a technology probe that encapsulates our framework. Our work highlights the potential for a hybrid LLM and optimization approach for iterative preference elicitation, and suggests design considerations for building systems that support human-system collaborative decision-making processes.

27 October 2024
Verified
Conference Paper
Open access
LLLindy Le
Lindy Le
The increasing prevalence of neurodivergence has led society to give greater recognition to the importance of neurodiversity. Yet societal perceptions of neurodivergence continue to be predominantly negative. Drawing on Critical Disability Studies, accessibility researchers have demonstrated how neuronormative assumptions dominate HCI. Despite their guidance, neurodivergent and disabled individuals are still marginalized in technology research. In particular, intersectional identities remain largely absent from HCI neurodivergence research. In this paper, I share my perspective as an outsider of the academic research community: I use critical autoethnography to analyze my experiences of coming to understand, accept, and value my neurodivergence within systems of power, privilege, and oppression. Using Data Feminism as an accessible and practical guide to intersectionality, I derive three tenets for reconceptualizing neurodivergence to be more inclusive of intersectional experiences: (1) neurodivergence is a functional difference, not a deficit; (2) neurodivergent disability is a moment of friction, not a static label; and (3) neurodivergence accessibility is a collaborative practice, not a one-sided solution. Then, I discuss the tenets in the context of existing HCI research, applying the same intersectional lens. Finally, I offer three suggestions for how accessibility research can apply these tenets in future work, to bridge the gap between accessibility theory and practice in HCI neurodivergence research.


16 February 2026 in IEEE Transactions on Affective Computing
Verified
Journal Article
Closed access
JSJina Suh
Jina Suh
LLLindy Le
Lindy Le
ESErfan Shayegani
Erfan Shayegani
GRGonzalo Ramos
Gonzalo Ramos
JAJudith Amores
Judith Amores
JHJavier Hernandez
Javier Hernandez
Empathy is increasingly recognized as a key factor in human–AI communication, yet conventional approaches to “digital empathy” often focus on simulating internal, human like emotional states while overlooking the inherently subjective, contextual, and relational facets of empathy as perceived by users. In this work, we propose a human-centered taxonomy that emphasizes observable empathic behaviors and introduce a new dataset, SENSE-7, of real-world conversations between information workers and Large Language Models (LLMs), which includes per-turn empathy annotations directly from the users, along with user characteristics, and contextual details, offering a more user-grounded representation of empathy. Analysis of 695 conversations from 109 participants reveals that empathy judgments are highly individualized, context-sensitive, and vulnerable to disruption when conversational continuity fails or user expectations go unmet. To promote further research, we provide a subset of 672 anonymized conversation and provide exploratory classification analysis, showing that an LLM-based classifier can recognize 5 levels of empathy with an encouraging average Spearman ρ = 0.369 and Accuracy = 0.487 over this set. Overall, our findings underscore the need for AI designs that dynamically tailor empathic behaviors to user contexts and goals, offering a roadmap for future research and practical development of socially attuned, human-centered artificial agents.

23 September 2026 in Proceedings of the ACM on Human-Computer Interaction
Verified
Journal Article
Open access
PKPranav Khadpe
Pranav Khadpe
LPLindsay Popowski
Lindsay Popowski
LLLindy Le
Lindy Le
KMKyzyl Monteiro
Kyzyl Monteiro
GKGeoff Kaufman
Geoff Kaufman
In an online community, prosocial actions ranging from sharing authentic opinions to intervening against misbehavior to contributing to collective action are often deterred by a first mover disadvantage: isolated individuals deciding whether or not to take the first action often cannot be sure whether others would welcome it, and respond in support. As a result, people may fail to surface an opinion even though it is privately held by many, refrain from publicly speaking up against misbehavior even though many privately think it is unacceptable, and fail to act in response to concerns that are not voiced publicly but are widespread. In this paper, we formalize how designers of online communities can lower these first-mover disadvantages through a design pattern that we call an action escrow —a mechanism where people deposit a socially risky action with an intermediary system that only executes the action if a prespecified trigger criterion is met. For example, an action escrow for encouraging authentic opinions might allow a user to place a comment into escrow with the instruction that it be posted publicly only if the escrow system receives similar comments from two other users. Although action escrows are not new—they feature in some existing systems and are inspired by traditional escrows in legal and economic scholarship—we formalize their scope, and utility for addressing persistent challenges in online communities. We explain the general design pattern, present design cases of implementations that apply the pattern to specific problems, describe the broader design space for action escrows, and outline opportunities for the application of escrows more generally, to address CSCW challenges.