Searcharxiv⌕ Search

arXiv subjects

Jason

Publications and source records attributed to Jason.

12 recordsLinked to original sources
arXiv2026

Beyond Her: Safety Dynamics in Role-play AI Companions

The film 'Her' pictured a future of love between humans and AI. That future has quietly emerged in the form of Role-play AI Companions (RACs), where emotionally responsive interactions blur the boundary between tool use and relational engagement. However, the safety implications remain poorly understood, as user experiences evolve over time through safety dynamics, spanning both emotional and risk behavioral dynamics, that can gradually shift interactions toward risk. In this paper, we investigate safety dynamics in RAC usage through a two-part mixed-methods study (Study I \& II). (1) Study I consists of semi-structured interviews (N = 16) to identify the key factors shaping these dynamics. We find that users' internalizing problems, the role personality adopted by the RAC, and risk interaction patterns jointly shape safety dynamics. Building on these insights, (2) Study II conducts a 14-day Ecological Momentary Assessment (N = 102) to examine how safety dynamics unfold in real-world usage. We identify distinct user profiles based on internalizing problems and show that interactions with RACs can produce short-term emotional relief while masking longer-term deterioration. Furthermore, vulnerable users exhibit more unstable risk behavioral patterns over time, making risk emergence less predictable and harder to mitigate with static safeguards. Our findings highlight the importance of modeling safety as a dynamic process rather than a static property. We conclude with three-layer design implications for next-generation AI companions, advocating for adaptive safeguards that can respond to evolving emotional and behavioral signals.

cs.CR↗
arXiv2026

Beyond Attack Success Rate: Examining Trigger Leakage in Vision-Language Agentic Systems

Vision-Language Agentic Systems (VLAS) connect visual perception to planning, tool use, and physical actions. This means backdoor-type triggers can propagate through both decision pipelines and their connected interfaces, thus making visual backdoors a system-level threat. Current evaluations on such backdoors focus on clean accuracy and attack success rate (ASR), metrics that capture whether a trigger works, but not whether an attack is actually "precise" -- i.e. whether it triggers hidden behaviors only when intended. In this work, we formalize the failure of trigger precision as "trigger leakage": inputs that are visually or semantically close to the intended trigger and therefore inadvertently activate the attacker-specified behavior. To quantify this leakage, we introduce Neighbor Leakage Rate (NLR). Our experiments show that at a 3% poisoning ratio, icon and text triggers remain robust to common visual transformations, but their neighboring variants leak heavily, with NLR reaching 0.996 (icon) and 0.944 (text). Using textual triggers as a controlled probe, we show that standard fine-tuning learns a broad activation region rather than an exact trigger condition, causing neighboring strings to invoke the malicious behavior even when the exact trigger is absent. Adding edit-distance-one hard-negative samples during training substantially narrows this activation region and reduces leakage, including in image-editing and embodied-manipulation workflows, where leaked triggers can propagate into executable programs and action sequences.

cs.CR↗
arXiv2026

PhoneWorld: Scaling Phone-Use Agent Environments

A central bottleneck for phone-use agents is that controllable, reproducible environments covering real mobile behavior are hard to build at scale. Existing mobile-agent benchmarks have made important progress on evaluation, but they do not by themselves provide a scalable way to construct many new phone-use environments. We present PhoneWorld, a reusable pipeline that converts real GUI trajectories and screenshots into controllable phone-use environments, executable tasks, automatic verifiers, and training rollouts. Rather than hand-building one mobile benchmark at a time, PhoneWorld uses real trajectories to recover which screens matter, how screens connect, which interactions must change environment state, and which user goals admit automatic verification. From these signals, it builds runnable mock Android apps backed by read-only app content and mutable state, then derives executable tasks, rule-based verifiers, and training rollouts from the same environments. In its current instantiation, PhoneWorld covers 34 apps across 16 domains, spanning common consumer mobile behaviors such as search, browsing, shopping, booking, media, and social interaction. Under a fixed training budget, replacing 10K steps from an auxiliary AndroidWorld corpus in an AndroidWorld-based baseline with broad PhoneWorld supervision improves all four evaluation benchmarks at once, raising HYMobileBench by 17.7 points, AndroidControl by 6.0 points, AndroidWorld by 14.7 points, and PhoneWorld by 52.5 points. We then study two additional scaling questions: increasing the amount of PhoneWorld supervision strongly improves PhoneWorld performance, and under a fixed PhoneWorld budget, expanding app coverage yields even larger gains. Overall, PhoneWorld shifts the focus from building one mobile benchmark at a time to scaling the supply of phone-use environments themselves.

cs.CL↗
arXiv2026

PPG-Based Heart Rate Accuracy in Diverse Populations: Investigating Inequities Across Body Composition and Skin Tones

Wearable devices are widely used for heart rate (HR) monitoring, yet their accuracy across diverse body compositions and skin tones remains uncertain. This study evaluated four wrist worn devices (Apple, Fitbit, Samsung, Garmin) in 58 Hispanic adults with Fitzpatrick skin types III to V during a cycling protocol alternating moderate (0.64 to 0.76 HRmax) and vigorous (0.77 to 0.95 HRmax) intensities. Criterion HR was obtained using a Polar H10 ECG, and accuracy was assessed using mean absolute error, mean absolute percentage error (MAPE), bias, and intraclass correlation coefficients. All devices showed significant deviation from criterion measures. Apple and Garmin demonstrated the lowest error, whereas Fitbit and Samsung exhibited greater inaccuracies. Higher BMI and darker skin tones were associated with increased MAPE. These biases disproportionately affect higher risk populations, underscoring the need for improved algorithms to ensure equitable health monitoring.

physics.med-ph↗
arXiv2025

Smart Fast Finish: Preventing Overdelivery via Daily Budget Pacing at DoorDash

We present a budget pacing feature called Smart Fast Finish (SFF). SFF builds upon the industry standard Fast Finish (FF) feature in budget pacing systems that depletes remaining advertising budget as quickly as possible towards the end of some fixed time period. SFF dynamically updates system parameters such as start time and throttle rate depending on historical ad-campaign data. SFF is currently in use at DoorDash, one of the largest delivery platforms in the US, and is part of its budget pacing system. We show via online budget-split experimentation data and offline simulations that SFF is a robust solution for overdelivery mitigation when pacing budget.

cs.GT↗
arXiv2022

GAN based Data Augmentation to Resolve Class Imbalance

The number of credit card fraud has been growing as technology grows and people can take advantage of it. Therefore, it is very important to implement a robust and effective method to detect such frauds. The machine learning algorithms are appropriate for these tasks since they try to maximize the accuracy of predictions and hence can be relied upon. However, there is an impending flaw where in machine learning models may not perform well due to the presence of an imbalance across classes distribution within the sample set. So, in many related tasks, the datasets have a very small number of observed fraud cases (sometimes around 1 percent positive fraud instances found). Therefore, this imbalance presence may impact any learning model's behavior by predicting all labels as the majority class, hence allowing no scope for generalization in the predictions made by the model. We trained Generative Adversarial Network(GAN) to generate a large number of convincing (and reliable) synthetic examples of the minority class that can be used to alleviate the class imbalance within the training set and hence generalize the learning of the data more effectively.

cs.LG↗
arXiv2019

Deep Reinforcement Learning for Personalized Search Story Recommendation

In recent years, \emph{search story}, a combined display with other organic channels, has become a major source of user traffic on platforms such as e-commerce search platforms, news feed platforms and web and image search platforms. The recommended search story guides a user to identify her own preference and personal intent, which subsequently influences the user's real-time and long-term search behavior. %With such an increased importance of search stories, As search stories become increasingly important, in this work, we study the problem of personalized search story recommendation within a search engine, which aims to suggest a search story relevant to both a search keyword and an individual user's interest. To address the challenge of modeling both immediate and future values of recommended search stories (i.e., cross-channel effect), for which conventional supervised learning framework is not applicable, we resort to a Markov decision process and propose a deep reinforcement learning architecture trained by both imitation learning and reinforcement learning. We empirically demonstrate the effectiveness of our proposed approach through extensive experiments on real-world data sets from JD.com.

cs.LG↗
arXiv2019

Measuring Average Treatment Effect from Heavy-tailed Data

Heavy-tailed metrics are common and often critical to product evaluation in the online world. While we may have samples large enough for Central Limit Theorem to kick in, experimentation is challenging due to the wide confidence interval of estimation. We demonstrate the pressure by running A/A simulations with customer spending data from a large-scale Ecommerce site. Solutions are then explored. On one front we address the heavy tail directly and highlight the often ignored nuances of winsorization. In particular, the legitimacy of false positive rate could be at risk. We are further inspired by the idea of robust statistics and introduce Huber regression as a better way to measure treatment effect. On another front covariates from pre-experiment period are exploited. Although they are independent to assignment and potentially explain the variation of response well, concerns are that models are learned against prediction error rather than the bias of parameter. We find the framework of orthogonal learning useful, matching not raw observations but residuals from two predictions, one towards the response and the other towards the assignment. Robust regression is readily integrated, together with cross-fitting. The final design is proven highly effective in driving down variance at the same time controlling bias. It is empowering our daily practice and hopefully can also benefit other applications in the industry.

stat.AP↗
arXiv2017

Inequalities, Preferences and Rankings in US Sports Coach Hiring Networks

Hiring a head coach of a college sports team is vital which will definitely have a great influence on the later development of the team. However, a lot of attention has been focused on each coach's individual features. A systematic and quantitative analysis of the whole coach hiring market is lacking. In a coach hiring network, the coaches are actually voting with their feet. It is interesting to analyze what factors are affecting the "footprint" left by those head coaches. In this paper, we collect more than 12,000 head coach hiring records in two different popular sports from the NCAA. Using network-based methods, we build the coach hiring network in the NCAA men's basketball and football. We find that: (1).the coach hiring network is of great inequality in coach production with a Gini coefficient close to 0.60. (2).coaches prefer to work within the same geographical region and the same division to their alma maters'. (3).the coach production rankings we calculated using network-based methods are generally correlated to the authoritative rankings, but also show disaccord in specific time period. The results provide us a novel view and better understanding of the coach hiring market in the NCAA and shed new light on the coach hiring system.

cs.SI↗
arXiv2017

Machine Learning Based Detection of Clickbait Posts in Social Media

Clickbait (headlines) make use of misleading titles that hide critical information from or exaggerate the content on the landing target pages to entice clicks. As clickbaits often use eye-catching wording to attract viewers, target contents are often of low quality. Clickbaits are especially widespread on social media such as Twitter, adversely impacting user experience by causing immense dissatisfaction. Hence, it has become increasingly important to put forward a widely applicable approach to identify and detect clickbaits. In this paper, we make use of a dataset from the clickbait challenge 2017 (clickbait-challenge.com) comprising of over 21,000 headlines/titles, each of which is annotated by at least five judgments from crowdsourcing on how clickbait it is. We attempt to build an effective computational clickbait detection model on this dataset. We first considered a total of 331 features, filtered out many features to avoid overfitting and improve the running time of learning, and eventually selected the 60 most important features for our final model. Using these features, Random Forest Regression achieved the following results: MSE=0.035 MSE, Accuracy=0.82, and F1-sore=0.61 on the clickbait class.

cs.CL↗
arXiv2010

Software Requirements Specification of the IUfA's UUIS -- a Team 2 COMP5541-W10 Project Approach

In the 52-page document, we describe our approach to the Software Requirements Specification of the IUfA's UUIS prototype. This includes the overall system description, functional requirements, non-functional requirements, use cases, the corresponding data dictionary for all entities involved, mock user interface (UI) design, and the overall projected cost estimate. The design specification of UUIS can be found in arXiv:1005.0665.

cs.SE↗