arXiv subjects

Lydia Chilton

Publications and source records attributed to Lydia Chilton.

17 recordsLinked to original sources

arXiv2026

FlowCheck: Helping End-Users Specify and Verify Intent in Vibe-Coded Web Apps

Vibe-coded applications often contain silent behavioral failures in which the interface appears functional even though user-visible information does not flow to the expected state or output. We introduce FlowCheck, a constraint language to specify these user-visible information flows directly through the application interface, where constraints can also be displayed and inspected without reading code, and are structured enough for reliable LLM generation. FlowCheck translates the constraints into deterministic CodeQL analyses, and we evaluate it across four applications generated via Claude Code, and compare with three coding models as bug-finding baselines. We find that FlowCheck correctly translates and flags all 30 of our injected constraint violations with no false positives. In contrast, frontier models (Claude Opus 4.7, DeepSeek V3, and Gemini Pro) showed significantly lower accuracy when prompted to find bugs in the same code, with none achieving full accuracy. This approach lets vibe coders state intent in terms of the interface they understand, and checks it deterministically against the code they do not.

cs.SE↗

arXiv2026

fog: Expressing Motion and Emotion through Function Composition of AI-Generated Code

Motion and emotion are core parts of intelligent, expressive behavior. In this paper, we introduce fog, a function composition framework for implementing and compose motion functions. We demonstrate how fog can be used to express motion and emotion in Heider-Simmel style animations. This code generation framework can help users generate functions for verbs, adverbs, gestures, and emotions to create an open-ended motion vocabulary. It is complemented by an animation editor that helps users refine motion through direct manipulation and dynamically generated UI. We evaluate our approach with a perceptual evaluation, where we test 452 fog-generated animations to see if people can recognize the semantic meaning of the motion. We find that fog's motion functions can be recognized at 68% accuracy, a 2.68x improvement over a chance baseline. In a mixed-methods user study with professionals and novices, we show that fog in interface form can support users with more rapid iteration, exploration, and control.

cs.HC↗

arXiv2026

PERSONAJUDGE: Simulating Individual Human Preference Judgments with Evaluator-Specific Demonstration Data

Large language models increasingly serve as judges in AI evaluation, but current approaches rely on consensus preferences that ignore individual evaluator variation. We propose a novel simulation approach that combines categorical judgments with evaluator-specific auxiliary data--retrospective reasoning traces and interface telemetry--to enable LLM-based simulation of individual evaluators via in-context learning. We conduct a systematic empirical study of this approach using multi-facet data from 32 trained annotators across 4,200 preference judgments in a 4 x 4 x 4 factorial design. Our key findings: (1) The simulation approach achieves up to 9.9 percentage point improvements over the Base Judge; (2) Reasoning traces provide the largest gains with higher collection efforts, while interface telemetry often hurts rather than helps performance despite being cheaper to collect. (3) Simulation difficulty is systematic, predicted by an evaluator's neutral usage (most clearly on Helpfulness) and divergence from consensus; the neutral-usage tendency--rather than simulatability itself--is the cross-task-stable property (r = 0.728). These results establish both the potential and limits of evaluator-specific auxiliary data for personalized evaluation, offering methodological insights for scaling individual aware AI assessment.

cs.HC↗

arXiv2026

Designing for the Moment: How One-Minute Interventions Fit or Falter Across Domains

This paper explores the design space for one-minute digital interventions that prompt immediate action without onboarding or sensing. By embracing Fogg's Behavior Model and four design principles informed by literature, the goal of these interventions was to provide triggers that encourage actions so simple that even people with low motivation would be willing to complete them. We examined the utility of these prompts by conducting a 14-day study with 22 participants interested in making small lifestyle improvements in at least one of three domains: physical activity, healthy eating, and mental well-being. When combined with insights drawn from participants' rewrites of our prompts, our findings suggest that intentional personalization through co-authorship could be a lightweight personalization mechanism that balances relevance with low friction.

cs.HC↗

arXiv2025

Micro-Health Interventions: Exploring Design Strategies for 1-Minute Interventions as a Gateway to Healthy Habits

One-minute behavior change interventions might seem too brief to matter. Could something so short really help people build healthier routines? This work explores this question through two studies examining how ultra-brief prompts might encourage meaningful actions in daily life. In a formative study, we explored how participants engaged with one-minute prompts across four domains: physical activity, eating, screen use, and mental well-being. This revealed two common design approaches: Immediate Action prompts (simple, directive tasks) and Reflection-First prompts (self-awareness before action). We then conducted a 14-day, within-subjects study comparing these two flows with 28 participants. Surprisingly, most participants did not notice differences in structure -- but responded positively when prompts felt timely, relevant, or emotionally supportive. Engagement was not shaped by flow type, but by content fit, tone, and momentary readiness. Participants also co-designed messages, favoring those with step-by-step guidance, personal meaning, or sensory detail. These results suggest that one-minute interventions, while easily dismissed, may serve as meaningful gateways into healthier routines -- if designed to feel helpful in the moment.

cs.HC↗

arXiv2025

OPeRA: A Dataset of Observation, Persona, Rationale, and Action for Evaluating LLMs on Human Online Shopping Behavior Simulation

Can large language models (LLMs) accurately simulate the next web action of a specific user? While LLMs have shown promising capabilities in generating ``believable'' human behaviors, evaluating their ability to mimic real user behaviors remains an open challenge, largely due to the lack of high-quality, publicly available datasets that capture both the observable actions and the internal reasoning of an actual human user. To address this gap, we introduce OPERA, a novel dataset of Observation, Persona, Rationale, and Action collected from real human participants during online shopping sessions. OPERA is the first public dataset that comprehensively captures: user personas, browser observations, fine-grained web actions, and self-reported just-in-time rationales. We developed both an online questionnaire and a custom browser plugin to gather this dataset with high fidelity. Using OPERA, we establish the first benchmark to evaluate how well current LLMs can predict a specific user's next action and rationale with a given persona and history. This dataset lays the groundwork for future research into LLM agents that aim to act as personalized digital twins for human.

cs.CL↗

arXiv2024

LogoMotion: Visually-Grounded Code Synthesis for Creating and Editing Animation

Creating animation takes time, effort, and technical expertise. To help novices with animation, we present LogoMotion, an AI code generation approach that helps users create semantically meaningful animation for logos. LogoMotion automatically generates animation code with a method called visually-grounded code synthesis and program repair. This method performs visual analysis, instantiates a design concept, and conducts visual checking to generate animation code. LogoMotion provides novices with code-connected AI editing widgets that help them edit the motion, grouping, and timing of their animation. In a comparison study on 276 animations, LogoMotion was found to produce more content-aware animation than an industry-leading tool. In a user evaluation (n=16) comparing against a prompt-only baseline, these code-connected widgets helped users edit animations with control, iteration, and creative expression.

cs.HC↗

arXiv2024

MoodSmith: Enabling Mood-Consistent Multimedia for AI-Generated Advocacy Campaigns

Emotion is vital to information and message processing, playing a key role in attitude formation. Consequently, creating a mood that evokes an emotional response is essential to any compelling piece of outreach communication. Many nonprofits and charities, despite having established messages, face challenges in creating advocacy campaign videos for social media. It requires significant creative and cognitive efforts to ensure that videos achieve the desired mood across multiple dimensions: script, visuals, and audio. We introduce MoodSmith, an AI-powered system that helps users explore mood possibilities for their message and create advocacy campaigns that are mood-consistent across dimensions. To achieve this, MoodSmith uses emotive language and plotlines for scripts, artistic style and color palette for visuals, and positivity and energy for audio. Our studies show that MoodSmith can effectively achieve a variety of moods, and the produced videos are consistent across media dimensions.

cs.HC↗

arXiv2024

Simulating Human Strategic Behavior: Comparing Single and Multi-agent LLMs

When creating policies, plans, or designs for people, it is challenging for designers to foresee all of the ways in which people may reason and behave. Recently, Large Language Models (LLMs) have been shown to be able to simulate human reasoning. We extend this work by measuring LLMs ability to simulate strategic reasoning in the ultimatum game, a classic economics bargaining experiment. Experimental evidence shows human strategic reasoning is complex; people will often choose to punish other players to enforce social norms even at personal expense. We test if LLMs can replicate this behavior in simulation, comparing two structures: single LLMs and multi-agent systems. We compare their abilities to (1) simulate human-like reasoning in the ultimatum game, (2) simulate two player personalities, greedy and fair, and (3) create robust strategies that are logically complete and consistent with personality. Our evaluation shows that multi-agent systems are more accurate than single LLMs (88 percent vs. 50 percent) in simulating human reasoning and actions for personality pairs. Thus, there is potential to use LLMs to simulate human strategic reasoning to help decision and policy-makers perform preliminary explorations of how people behave in systems.

cs.HC↗

arXiv2023

STORYWARS: A Dataset and Instruction Tuning Baselines for Collaborative Story Understanding and Generation

Collaborative stories, which are texts created through the collaborative efforts of multiple authors with different writing styles and intentions, pose unique challenges for NLP models. Understanding and generating such stories remains an underexplored area due to the lack of open-domain corpora. To address this, we introduce STORYWARS, a new dataset of over 40,000 collaborative stories written by 9,400 different authors from an online platform. We design 12 task types, comprising 7 understanding and 5 generation task types, on STORYWARS, deriving 101 diverse story-related tasks in total as a multi-task benchmark covering all fully-supervised, few-shot, and zero-shot scenarios. Furthermore, we present our instruction-tuned model, INSTRUCTSTORY, for the story tasks showing that instruction tuning, in addition to achieving superior results in zero-shot and few-shot scenarios, can also obtain the best performance on the fully-supervised tasks in STORYWARS, establishing strong multi-task benchmark performances on STORYWARS.

cs.CL↗

arXiv2023

Generative Disco: Text-to-Video Generation for Music Visualization

Visuals can enhance our experience of music, owing to the way they can amplify the emotions and messages conveyed within it. However, creating music visualization is a complex, time-consuming, and resource-intensive process. We introduce Generative Disco, a generative AI system that helps generate music visualizations with large language models and text-to-video generation. The system helps users visualize music in intervals by finding prompts to describe the images that intervals start and end on and interpolating between them to the beat of the music. We introduce design patterns for improving these generated videos: transitions, which express shifts in color, time, subject, or style, and holds, which help focus the video on subjects. A study with professionals showed that transitions and holds were a highly expressive framework that enabled them to build coherent visual narratives. We conclude on the generalizability of these patterns and the potential of generated video for creative professionals.

cs.HC↗

arXiv2022

SafeText: A Benchmark for Exploring Physical Safety in Language Models

Understanding what constitutes safe text is an important issue in natural language processing and can often prevent the deployment of models deemed harmful and unsafe. One such type of safety that has been scarcely studied is commonsense physical safety, i.e. text that is not explicitly violent and requires additional commonsense knowledge to comprehend that it leads to physical harm. We create the first benchmark dataset, SafeText, comprising real-life scenarios with paired safe and physically unsafe pieces of advice. We utilize SafeText to empirically study commonsense physical safety across various models designed for text generation and commonsense reasoning tasks. We find that state-of-the-art large language models are susceptible to the generation of unsafe text and have difficulty rejecting unsafe advice. As a result, we argue for further studies of safety and the assessment of commonsense physical safety in models before release.

cs.CL↗

arXiv2022

Opal: Multimodal Image Generation for News Illustration

Advances in multimodal AI have presented people with powerful ways to create images from text. Recent work has shown that text-to-image generations are able to represent a broad range of subjects and artistic styles. However, finding the right visual language for text prompts is difficult. In this paper, we address this challenge with Opal, a system that produces text-to-image generations for news illustration. Given an article, Opal guides users through a structured search for visual concepts and provides a pipeline allowing users to generate illustrations based on an article's tone, keywords, and related artistic styles. Our evaluation shows that Opal efficiently generates diverse sets of news illustrations, visual assets, and concept ideas. Users with Opal generated two times more usable results than users without. We discuss how structured exploration can help users better understand the capabilities of human AI co-creative systems.

cs.HC↗

arXiv2021

Lightweight Decoding Strategies for Increasing Specificity

Language models are known to produce vague and generic outputs. We propose two unsupervised decoding strategies based on either word-frequency or point-wise mutual information to increase the specificity of any model that outputs a probability distribution over its vocabulary at generation time. We test the strategies in a prompt completion task; with human evaluations, we find that both strategies increase the specificity of outputs with only modest decreases in sensibility. We also briefly present a summarization use case, where these strategies can produce more specific summaries.

cs.CL↗

arXiv2021

Hierarchical Summarization for Longform Spoken Dialog

Every day we are surrounded by spoken dialog. This medium delivers rich diverse streams of information auditorily; however, systematically understanding dialog can often be non-trivial. Despite the pervasiveness of spoken dialog, automated speech understanding and quality information extraction remains markedly poor, especially when compared to written prose. Furthermore, compared to understanding text, auditory communication poses many additional challenges such as speaker disfluencies, informal prose styles, and lack of structure. These concerns all demonstrate the need for a distinctly speech tailored interactive system to help users understand and navigate the spoken language domain. While individual automatic speech recognition (ASR) and text summarization methods already exist, they are imperfect technologies; neither consider user purpose and intent nor address spoken language induced complications. Consequently, we design a two stage ASR and text summarization pipeline and propose a set of semantic segmentation and merging algorithms to resolve these speech modeling challenges. Our system enables users to easily browse and navigate content as well as recover from errors in these underlying technologies. Finally, we present an evaluation of the system which highlights user preference for hierarchical summarization as a tool to quickly skim audio and identify content of interest to the user.

cs.CL↗

arXiv2019

Low-Level Linguistic Controls for Style Transfer and Content Preservation

Despite the success of style transfer in image processing, it has seen limited progress in natural language generation. Part of the problem is that content is not as easily decoupled from style in the text domain. Curiously, in the field of stylometry, content does not figure prominently in practical methods of discriminating stylistic elements, such as authorship and genre. Rather, syntax and function words are the most salient features. Drawing on this work, we model style as a suite of low-level linguistic controls, such as frequency of pronouns, prepositions, and subordinate clause constructions. We train a neural encoder-decoder model to reconstruct reference sentences given only content words and the setting of the controls. We perform style transfer by keeping the content words fixed while adjusting the controls to be indicative of another style. In experiments, we show that the model reliably responds to the linguistic controls and perform both automatic and manual evaluations on style transfer. We find we can fool a style classifier 84% of the time, and that our model produces highly diverse and stylistically distinctive outputs. This work introduces a formal, extendable model of style that can add control to any neural text generation system.

cs.CL↗

arXiv2010

The Labor Economics of Paid Crowdsourcing

Crowdsourcing is a form of "peer production" in which work traditionally performed by an employee is outsourced to an "undefined, generally large group of people in the form of an open call." We present a model of workers supplying labor to paid crowdsourcing projects. We also introduce a novel method for estimating a worker's reservation wage--the smallest wage a worker is willing to accept for a task and the key parameter in our labor supply model. It shows that the reservation wages of a sample of workers from Amazon's Mechanical Turk (AMT) are approximately log normally distributed, with a median wage of $1.38/hour. At the median wage, the point elasticity of extensive labor supply is 0.43. We discuss how to use our calibrated model to make predictions in applied work. Two experimental tests of the model show that many workers respond rationally to offered incentives. However, a non-trivial fraction of subjects appear to set earnings targets. These "target earners" consider not just the offered wage--which is what the rational model predicts--but also their proximity to earnings goals. Interestingly, a number of workers clearly prefer earning total amounts evenly divisible by 5, presumably because these amounts make good targets.

cs.HC↗