Searcharxiv⌕ Search

arXiv subjects

Sangwook Lee

Publications and source records attributed to Sangwook Lee.

At least 19 recordsLinked to original sources
arXiv2026

The Help Ladder: Skill-Adaptive Peer Scaffolding for Real-Time Collaborative Programming

Collaborative programming is a widely adopted classroom activity to encourage peer scaffolding, yet real-time collaboration often breaks down into parallel individual work with minimal interaction. Our formative studies reveal that even when students want to collaborate, they are held back by the effort required to understand a teammate's entire problem at once. We present Canary, a system that supports peer scaffolding by breaking down programming obstacles into smaller steps tailored to a student's skill level. Canary alerts potential helpers to specific places where they can start, using AI to turn complex problems into a step-by-step ladder that starts with easy fixes before moving toward harder logic. By providing this gradual ramp-up, Canary enables students to make quick contributions and progressively work toward solving their peers' problems. Our evaluation shows that this staged approach makes helping feel less overwhelming, leading to more frequent and effective collaboration among students.

cs.HC↗
arXiv2026

NIRVANA: A Comprehensive Dataset for Reproducing How Students Use Generative AI for Essay Writing

With the rapid adoption of AI writing assistants in education, educators and researchers need empirical evidence to understand the impact on student writing and inform effective pedagogical design. Despite widespread use, we lack systematic understanding of how students engage with these tools during authentic writing tasks: when they seek assistance, what they ask, and how they incorporate AI-generated content into their essays. This gap limits evidence-based policy development and rigorous evaluation of generative AI's learning effects. To address this gap, we introduce NIRVANA, a dataset capturing how university students use generative AI while writing an analytical essay. The dataset includes 77 students who completed an essay task with access to ChatGPT, recording keystroke-level writing behavior, full ChatGPT conversation histories, and all text copied from ChatGPT, enabling a complete reconstruction of the writing process and revealing how AI assistance shapes student work. Our analysis identifies key behavioral patterns, including variation in ChatGPT query frequency and its relationship to essay characteristics such as length and readability. We identify four writing profiles based on students' contribution and revision patterns: Lead Authors, Collaborators, Drafters, and Vibe Writers. To support deeper investigation, we developed a replay interface that reconstructs the writing process; qualitative analysis of sampled replays demonstrates how this tool enables systematic examination of student-AI interactions.

cs.HC↗
arXiv2026

MAESTRO: Adapting GUIs and Guiding Navigation with User Preferences in Conversational Agents with GUIs

Modern task-oriented chatbots present GUI elements alongside natural-language dialogue, yet the agent's role has largely been limited to interpreting natural-language input as GUI actions and following a linear workflow. In preference-driven, multi-step tasks such as booking a flight or reserving a restaurant, earlier choices constrain later options and may force users to restart from scratch. User preferences serve as the key criteria for these decisions, yet existing agents do not systematically leverage them. We present MAESTRO, which extends the agent's role from execution to decision support. MAESTRO maintains a shared preference memory that extracts hard and soft preferences from natural-language utterances and provides two mechanisms. Preference-Grounded GUI Adaptation applies in-place operators (augment, sort, filter, and highlight) to the existing GUI according to preference strength, supporting comparison among options. Preference-Guided Workflow Navigation detects conflicts between preferences and available options, proposes backtracking, and records failed paths to avoid revisiting dead ends. Through a controlled experiment (N=33), we demonstrated that MAESTRO improved decision quality in movie ticketing: final bookings left fewer hard preferences unmet, and users made fewer selections that violated their stated preferences during the process than in the baseline condition, although task success rate and completion time did not differ significantly. In addition, we showed that using MAESTRO in voice mode can increase users' active engagement as well as their mental burden, revealing the nuanced tension in agentic interaction design for conversational agents with a GUI.

cs.HC↗
arXiv2025

Closed-string mirror symmetry for dimer models

For all punctured Riemann surfaces arising as mirror curves of toric Calabi--Yau threefolds, we show that their symplectic cohomology is isomorphic to the compactly supported Hochschild cohomology of the noncommutative Landau--Ginzburg model defined on the NCCR of the associated toric Gorenstein singularities. This mirror correspondence is established by analyzing the closed-open map with boundaries on certain combinatorially defined immersed Lagrangians in the Riemann surface, yielding a ring isomorphism. We give a detailed examination of the properties of this isomorphism, emphasizing its relationship to the singularity structure.

math.SG↗
arXiv2025

CHOIR: A Chatbot-mediated Organizational Memory Leveraging Communication in University Research Labs

University research labs often rely on chat-based platforms for communication and project management, where valuable knowledge surfaces but is easily lost in message streams. Documentation can preserve knowledge, but it requires ongoing maintenance and is challenging to navigate. Drawing on formative interviews that revealed organizational memory challenges in labs, we designed CHOIR, an LLM-based chatbot that supports organizational memory through four key functions: document-grounded Q\&A, Q\&A sharing for follow-up discussion, knowledge extraction from conversations, and AI-assisted document updates. We deployed CHOIR in four research labs for one month (n=21), where the lab members asked 107 questions and lab directors updated documents 38 times in the organizational memory. Our findings reveal a privacy-awareness tension: questions were asked privately, limiting directors' visibility into documentation gaps. Students often avoided contribution due to challenges in generalizing personal experiences into universal documentation. We contribute design implications for privacy-preserving awareness and supporting context-specific knowledge documentation.

cs.HC↗
arXiv2025

CHOIR: Chat-based Helper for Organizational Intelligence Repository

Modern organizations frequently rely on chat-based platforms (e.g., Slack, Microsoft Teams, and Discord) for day-to-day communication and decision-making. As conversations evolve, organizational knowledge can get buried, prompting repeated searches and discussions. While maintaining shared documents, such as Wiki articles for the organization, offers a partial solution, it requires manual and timely efforts to keep it up to date, and it may not effectively preserve the social and contextual aspect of prior discussions. Moreover, reaching a consensus on document updates with relevant stakeholders can be time-consuming and complex. To address these challenges, we introduce CHOIR (Chat-based Helper for Organizational Intelligence Repository), a chatbot that integrates seamlessly with chat platforms. CHOIR automatically identifies and proposes edits to related documents, initiates discussions with relevant team members, and preserves contextual revision histories. By embedding knowledge management directly into chat environments and leveraging LLMs, CHOIR simplifies manual updates and supports consensus-driven editing based on maintained context with revision histories. We plan to design, deploy, and evaluate CHOIR in the context of maintaining an organizational memory for a research lab. We describe the chatbot's motivation, design, and early implementation to show how CHOIR streamlines collaborative document management.

cs.HC↗
arXiv2024

Kodaira-Spencer maps for elliptic orbispheres as isomorphisms of Frobenius algebras

Given a mirror pair of a symplectic manifold $X$ and a Landau-Ginzburg potential $W$, we are interested in the problem whether the quantum cohomology of $X$ and the Jacobian algebra of $W$ are isomorphic. Since those can be equipped with Frobenius algebra structures, we might ask whether they are isomorphic as Frobenius algebras. We show that the Kodaira-Spencer map gives a Frobenius algebra isomorphism for elliptic orbispheres, under the Floer theoretic modification of the residue pairing.

math.SG↗
arXiv2024

Orbifold Kodaira-Spencer maps and closed-string mirror symmetry for punctured Riemann surfaces

When a Weinstein manifold admits an action of a finite abelian group, we propose its mirror construction following the equivariant TQFT-type construction, and obtain as a mirror the orbifolding of the mirror of the quotient with respect to the induced dual group action. As an application, we construct an orbifold Landau-Ginzburg mirror of a punctured Riemann surface given as an abelian cover of the pair-of-pants, and prove its closed-string mirror symmetry using the (part of) closed-open map twisted by the dual group action.

math.SG↗
arXiv2022

ModSandbox: Facilitating Online Community Moderation Through Error Prediction and Improvement of Automated Rules

Despite the common use of rule-based tools for online content moderation, human moderators still spend a lot of time monitoring them to ensure that they work as intended. Based on surveys and interviews with Reddit moderators who use AutoModerator, we identified the main challenges in reducing false positives and false negatives of automated rules: not being able to estimate the actual effect of a rule in advance and having difficulty figuring out how the rules should be updated. To address these issues, we built ModSandbox, a novel virtual sandbox system that detects possible false positives and false negatives of a rule to be improved and visualizes which part of the rule is causing issues. We conducted a user study with online content moderators, finding that ModSandbox can support quickly finding possible false positives and false negatives of automated rules and guide moderators to update those to reduce future errors.

cs.HC↗
arXiv2022

LV-Linker: Supporting Linked Exploration of Phone Usage Log Data and Screen Video Data

Prior HCI studies often analyzed smartphone app usage data for usability and user experience research purposes. App usage videos are often collected by a screen recording app in order to better analyze the app usage behaviors (e.g., app usage time, screen transition, and notification handling). However, it is difficult to analyze app usage videos along with multiple user interaction stream data. When the length of a video is long, data analysis tends to take a long time due to the volume of user interaction data. This is even more difficult for novice researchers due to a lack of data analysis experience. In this paper, we propose LV-Linker (Log and Video Linker), a visualization tool that helps researchers quickly explore the app usage log and video data by linking multiple time series log data with the video data. We conducted a preliminary user study with eight participants to evaluate the benefits of linking, by measuring task completion time, helpfulness, and subjective task workload. Our results showed that offering a linking feature significantly lowers the task completion time and task workload.

cs.HC↗
arXiv2022

Twisted Jacobian algebras as endomorphism algebras of equivariant matrix factorizations

Given a polynomial $W$ with an isolated singularity, we can consider the Jacobian ring as an invariant of the singularity. If in addition we have a group action on the polynomial ring with $W$ fixed, we are led to consider the twisted Jacobian ring which reflects the equivariant structure as well. Our main result is to show that the twisted Jacobian ring is isomorphic to an endomorphism ring of the "twisted diagonal" matrix factorization. As an application, we suggest a way to investigate Floer theory of Lagrangian submanifolds which represent homological mirror functors.

math.AG↗
arXiv2020

Kodaira-Spencer map, Lagrangian Floer theory and orbifold Jacobian algebras

A version of mirror symmetry predicts a ring isomorphism between quantum cohomology of a symplectic manifold and Jacobian algebra of the Landau-Ginzburg mirror, and for toric manifolds Fukaya-Oh-Ohta-Ono constructed such a map called Kodaira-Spencer map using Lagrangian Floer theory. We discuss a general construction of Kodaira-Spencer ring homomorphism when LG mirror potential $W$ is given by $J$-holomorphic discs with boundary on a Lagrangian $L$: we find an $A_{\infty}$-algebra $\mathcal{B}$ whose $m_1$-complex is a Koszul complex for $W$ under mild assumptions on $L$. Closed-open map gives a ring homomorphism from quantum cohomology to cohomology algebra of $\mathcal{B}$ which is Jacobian algebra of $W$. We also construct an equivariant version for orbifold LG mirror $(W,H)$. We construct a Kodaira-Spencer map from quantum cohomology to another $A_{\infty}$-algebra $(\mathcal{B}\rtimes H)^H$ whose cohomology algebra is isomorphic to the orbifold Jacobian algebra of $(W,H)$ under an assumption. For the $2$-torus whose mirror is an orbifold LG model given by Fermat cubic with a $\mathbb{Z}/3$-action, we compute an explicit Kodaira-Spencer isomorphism.

math.SG↗
arXiv2020

Noncommutative homological mirror symmetry of elliptic curves

We prove an equivalence of two A-infinity functors, via Orlov's Landau-Ginzburg/Calabi-Yau correspondence. One is the Polishchuk-Zaslow's mirror symmetry functor of elliptic curves, and the other is a localized mirror functor from the Fukaya category of the 2-torus to a category of noncommutative matrix factorizations. As a corollary we prove that the noncommutative mirror functor realizes homological mirror symmetry for any translation parameter $t$.

math.SG↗
arXiv2020

Pairings in mirror symmetry between a symplectic manifold and a Landau-Ginzburg $B$-model

We find a relation between Lagrangian Floer pairing of a symplectic manifold and Kapustin-Li pairing of the mirror Landau-Ginzburg model under localized mirror functor. They are conformally equivalent with an interesting conformal factor $(vol^{Floer}/vol)^2$, which can be described as a ratio of Lagrangian Floer volume class and classical volume class. For this purpose, we introduce $B$-invariant of Lagrangian Floer cohomology with values in Jacobian ring of the mirror potential function. And we prove what we call a multi-crescent Cardy identity under certain conditions, which is a generalized form of Cardy identity. As an application, we discuss the case of general toric manifold, and the relation to the work of Fukaya-Oh-Ohta-Ono and their $Z$-invariant. Also, we compute the conformal factor $(vol^{Floer}/vol)^2$ for the elliptic curve quotient $\mathbb{P}^1_{3,3,3}$, which is expected to be related to the choice of a primitive form.

math.SG↗
arXiv2019

Drivers Drowsiness Detection using Condition-Adaptive Representation Learning Framework

We propose a condition-adaptive representation learning framework for the driver drowsiness detection based on 3D-deep convolutional neural network. The proposed framework consists of four models: spatio-temporal representation learning, scene condition understanding, feature fusion, and drowsiness detection. The spatio-temporal representation learning extracts features that can describe motions and appearances in video simultaneously. The scene condition understanding classifies the scene conditions related to various conditions about the drivers and driving situations such as statuses of wearing glasses, illumination condition of driving, and motion of facial elements such as head, eye, and mouth. The feature fusion generates a condition-adaptive representation using two features extracted from above models. The detection model recognizes drivers drowsiness status using the condition-adaptive representation. The condition-adaptive representation learning framework can extract more discriminative features focusing on each scene condition than the general representation so that the drowsiness detection method can provide more accurate results for the various driving situations. The proposed framework is evaluated with the NTHU Drowsy Driver Detection video dataset. The experimental results show that our framework outperforms the existing drowsiness detection methods based on visual analysis.

cs.CV↗
arXiv2019

Comparison of mirror functors of elliptic curves via LG/CY correspondence

Polishchuk-Zaslow explained the homological mirror symmetry between Fukaya category of symplectic torus and the derived category of coherent sheaves of elliptic curves via Lagrangian torus fibration. Recently, Cho-Hong-Lau found another proof of homological mirror symmetry using localized mirror functor, whose target category is given by graded matrix factorizations. We find an explicit relation between these two approaches.

math.SG↗