arXiv subjects

Taewook Kim

Publications and source records attributed to Taewook Kim.

13 recordsLinked to original sources

arXiv2026

Personagram: Bridging Personas and Product Design for Creative Ideation with Multimodal LLMs

Product designers often begin their design process with handcrafted personas. While personas are intended to ground design decisions in consumer preferences, they often fall short in practice by remaining abstract, expensive to produce, and difficult to translate into actionable design features. As a result, personas risk serving as static reference points rather than tools that actively shape design outcomes. To address these challenges, we built Personagram, an interactive system powered by multimodal large language models (MLLMs) that helps designers explore detailed census-based personas, extract product features inferred from persona attributes, and recombine them for specific customer segments. In a study with 12 professional designers, we show that Personagram facilitates more actionable ideation workflows by structuring multimodal thinking from persona attributes to product design features, achieving higher engagement with personas, perceived transparency, and satisfaction compared to a chat-based baseline. We discuss implications of integrating AI-generated personas into product design workflows.

cs.HC↗

arXiv2025

LLMs Behind the Scenes: Enabling Narrative Scene Illustration

Generative AI has established the opportunity to readily transform content from one medium to another. This capability is especially powerful for storytelling, where visual illustrations can illuminate a story originally expressed in text. In this paper, we focus on the task of narrative scene illustration, which involves automatically generating an image depicting a scene in a story. Motivated by recent progress on text-to-image models, we consider a pipeline that uses LLMs as an interface for prompting text-to-image models to generate scene illustrations given raw story text. We apply variations of this pipeline to a prominent story corpus in order to synthesize illustrations for scenes in these stories. We conduct a human annotation task to obtain pairwise quality judgments for these illustrations. The outcome of this process is the SceneIllustrations dataset, which we release as a new resource for future work on cross-modal narrative transformation. Through our analysis of this dataset and experiments modeling illustration quality, we demonstrate that LLMs can effectively verbalize scene knowledge implicitly evoked by story text. Moreover, this capability is impactful for generating and evaluating illustrations.

cs.CL↗

arXiv2025

Scaffolding Recursive Divergence and Convergence in Story Ideation

Human creative ideation involves both exploration of diverse ideas (divergence) and selective synthesis of explored ideas into coherent combinations (convergence). While processes of divergence and convergence are often interleaved and nested, existing AI-powered creativity support tools (CSTs) lack support for sophisticated orchestration of divergence and convergence. We present Reverger, an AI-powered CST that helps users ideate variations of conceptual directions for modifying a story by scaffolding flexible iteration between divergence and convergence. For divergence, our tool enables recursive exploration of alternative high-level directions for modifying a specific part of the original story. For convergence, it allows users to collect explored high-level directions and synthesize them into concrete variations. Users can then iterate between divergence and convergence until they find a satisfactory outcome. A within-subject study revealed that Reverger permitted participants to explore more unexpected and diverse high-level directions than a comparable baseline. Reverger users also felt that they had more fine-grained control and discovered more effort-worthy outcomes.

cs.HC↗

arXiv2025

Drama Llama: An LLM-Powered Storylets Framework for Authorable Responsiveness in Interactive Narrative

In this paper, we present Drama Llama, an LLM-powered storylets framework that supports the authoring of responsive, open-ended interactive stories. DL combines the structural benefits of storylet-based systems with the generative capabilities of large language models, enabling authors to create responsive interactive narratives while maintaining narrative control. Rather than crafting complex logical preconditions in a general-purpose or domain-specific programming language, authors define triggers in natural language that fire at appropriate moments in the story. Through a preliminary authoring study with six content authors, we present initial evidence that DL can generate coherent and meaningful narratives with believable character interactions. This work suggests directions for hybrid approaches that enhance authorial control while supporting emergent narrative generation through LLMs.

cs.HC↗

arXiv2024

Conditional Text-to-Image Generation with Reference Guidance

Text-to-image diffusion models have demonstrated tremendous success in synthesizing visually stunning images given textual instructions. Despite remarkable progress in creating high-fidelity visuals, text-to-image models can still struggle with precisely rendering subjects, such as text spelling. To address this challenge, this paper explores using additional conditions of an image that provides visual guidance of the particular subjects for diffusion models to generate. In addition, this reference condition empowers the model to be conditioned in ways that the vocabularies of the text tokenizer cannot adequately represent, and further extends the model's generalization to novel capabilities such as generating non-English text spellings. We develop several small-scale expert plugins that efficiently endow a Stable Diffusion model with the capability to take different references. Each plugin is trained with auxiliary networks and loss functions customized for applications such as English scene-text generation, multi-lingual scene-text generation, and logo-image generation. Our expert plugins demonstrate superior results than the existing methods on all tasks, each containing only 28.55M trainable parameters.

cs.CV↗

arXiv2024

Steering AI-Driven Personalization of Scientific Text for General Audiences

Digital media platforms (e.g., science blogs) offer opportunities to communicate scientific content to general audiences at scale. However, these audiences vary in their scientific expertise, literacy levels, and personal backgrounds, making effective science communication challenging. To address this challenge, we designed TranSlider, an AI-powered tool that generates personalized translations of scientific text based on individual user profiles (e.g., hobbies, location, and education). Our tool features an interactive slider that allows users to steer the degree of personalization from 0 (weakly relatable) to 100 (strongly relatable), leveraging LLMs to generate the translations with chosen degrees. Through an exploratory study with 15 participants, we investigated both the utility of these AI-personalized translations and how interactive reading features influenced users' understanding and reading experiences. We found that participants who preferred higher degrees of personalization appreciated the relatable and contextual translations, while those who preferred lower degrees valued concise translations with subtle contextualization. Furthermore, participants reported the compounding effect of multiple translations on their understanding of scientific content. Drawing on these findings, we discuss several implications for facilitating science communication and designing steerable interfaces to support human-AI alignment.

cs.HC↗

arXiv2024

Learning to Customize Text-to-Image Diffusion In Diverse Context

Most text-to-image customization techniques fine-tune models on a small set of \emph{personal concept} images captured in minimal contexts. This often results in the model becoming overfitted to these training images and unable to generalize to new contexts in future text prompts. Existing customization methods are built on the success of effectively representing personal concepts as textual embeddings. Thus, in this work, we resort to diversifying the context of these personal concepts \emph{solely} within the textual space by simply creating a contextually rich set of text prompts, together with a widely used self-supervised learning objective. Surprisingly, this straightforward and cost-effective method significantly improves semantic alignment in the textual space, and this effect further extends to the image space, resulting in higher prompt fidelity for generated images. Additionally, our approach does not require any architectural modifications, making it highly compatible with existing text-to-image customization methods. We demonstrate the broad applicability of our approach by combining it with four different baseline methods, achieving notable CLIP score improvements.

cs.CV↗

arXiv2024

MentalImager: Exploring Generative Images for Assisting Support-Seekers' Self-Disclosure in Online Mental Health Communities

Support-seekers' self-disclosure of their suffering experiences, thoughts, and feelings in the post can help them get needed peer support in online mental health communities (OMHCs). However, such mental health self-disclosure could be challenging. Images can facilitate the manifestation of relevant experiences and feelings in the text; yet, relevant images are not always available. In this paper, we present a technical prototype named MentalImager and validate in a human evaluation study that it can generate topical- and emotional-relevant images based on the seekers' drafted posts or specified keywords. Two user studies demonstrate that MentalImager not only improves seekers' satisfaction with their self-disclosure in their posts but also invokes support-providers' empathy for the seekers and willingness to offer help. Such improvements are credited to the generated images, which help seekers express their emotions and inspire them to add more details about their experiences and feelings. We report concerns on MentalImager and discuss insights for supporting self-disclosure in OMHCs.

cs.HC↗

arXiv2024

Authors' Values and Attitudes Towards AI-bridged Scalable Personalization of Creative Language Arts

Generative AI has the potential to create a new form of interactive media: AI-bridged creative language arts (CLA), which bridge the author and audience by personalizing the author's vision to the audience's context and taste at scale. However, it is unclear what the authors' values and attitudes would be regarding AI-bridged CLA. To identify these values and attitudes, we conducted an interview study with 18 authors across eight genres (e.g., poetry, comics) by presenting speculative but realistic AI-bridged CLA scenarios. We identified three benefits derived from the dynamics between author, artifact, and audience: those that 1) authors get from the process, 2) audiences get from the artifact, and 3) authors get from the audience. We found how AI-bridged CLA would either promote or reduce these benefits, along with authors' concerns. We hope our investigation hints at how AI can provide intriguing experiences to CLA audiences while promoting authors' values.

cs.HC↗

arXiv2023

Opportunities in Mental Health Support for Informal Dementia Caregivers Suffering from Verbal Agitation

People with dementia (PwD) often present verbal agitation such as cursing, screaming, and persistently complaining. Verbal agitation can impose mental distress on informal caregivers (e.g., family, friends), which may cause severe mental illnesses, such as depression and anxiety disorders. To improve informal caregivers' mental health, we explore design opportunities by interviewing 11 informal caregivers suffering from verbal agitation of PwD. In particular, we first characterize how the predictability of verbal agitation impacts informal caregivers' mental health and how caregivers' coping strategies vary before, during, and after verbal agitation. Based on our findings, we propose design opportunities to improve the mental health of informal caregivers suffering from verbal agitation: distracting PwD (in-situ support; before), prompting just-in-time maneuvers (information support; during), and comfort and education (social & information support; after). We discuss our reflections on cultural disparities between participants. Our work envisions a broader design space for supporting informal caregivers' well-being and describes when and how that support could be provided.

cs.HC↗

arXiv2022

RLens: A Computer-aided Visualization System for Supporting Reflection on Language Learning under Distributed Tutorship

With the rise of the gig economy, online language tutoring platforms are becoming increasingly popular. These platforms provide temporary and flexible jobs for native speakers as tutors and allow language learners to have one-on-one speaking practices on demand, on which learners occasionally practice the language with different tutors. With such distributed tutorship, learners can hold flexible schedules and receive diverse feedback. However, learners face challenges in consistently tracking their learning progress because different tutors provide feedback from diverse standards and perspectives, and hardly refer to learners' previous experiences with other tutors. We present RLens, a visualization system for facilitating learners' learning progress reflection by grouping different tutors' feedback, tracking how each feedback type has been addressed across learning sessions, and visualizing the learning progress. We validate our design through a between-subjects study with 40 real-world learners. Results show that learners can successfully analyze their progress and common language issues under distributed tutorship with RLens, while most learners using the baseline interface had difficulty achieving reflection tasks. We further discuss design considerations of computer-aided systems for supporting learning under distributed tutorship.

cs.HC↗

arXiv2020

Suppressing Spoof-irrelevant Factors for Domain-agnostic Face Anti-spoofing

Face anti-spoofing aims to prevent false authentications of face recognition systems by distinguishing whether an image is originated from a human face or a spoof medium. We propose a novel method called Doubly Adversarial Suppression Network (DASN) for domain-agnostic face anti-spoofing; DASN improves the generalization ability to unseen domains by learning to effectively suppress spoof-irrelevant factors (SiFs) (e.g., camera sensors, illuminations). To achieve our goal, we introduce two types of adversarial learning schemes. In the first adversarial learning scheme, multiple SiFs are suppressed by deploying multiple discrimination heads that are trained against an encoder. In the second adversarial learning scheme, each of the discrimination heads is also adversarially trained to suppress a spoof factor, and the group of the secondary spoof classifier and the encoder aims to intensify the spoof factor by overcoming the suppression. We evaluate the proposed method on four public benchmark datasets, and achieve remarkable evaluation results. The results demonstrate the effectiveness of the proposed method.

cs.CV↗

arXiv2018

BAN: Focusing on Boundary Context for Object Detection

Visual context is one of the important clue for object detection and the context information for boundaries of an object is especially valuable. We propose a boundary aware network (BAN) designed to exploit the visual contexts including boundary information and surroundings, named boundary context, and define three types of the boundary contexts: side, vertex and in/out-boundary context. Our BAN consists of 10 sub-networks for the area belonging to the boundary contexts. The detection head of BAN is defined as an ensemble of these sub-networks with different contributions depending on the sub-problem of detection. To verify our method, we visualize the activation of the sub-networks according to the boundary contexts and empirically show that the sub-networks contribute more to the related sub-problem in detection. We evaluate our method on PASCAL VOC detection benchmark and MS COCO dataset. The proposed method achieves the mean Average Precision (mAP) of 83.4% on PASCAL VOC and 36.9% on MS COCO. BAN allows the convolution network to provide an additional source of contexts for detection and selectively focus on the more important contexts, and it can be generally applied to many other detection methods as well to enhance the accuracy in detection.

cs.CV↗