arXiv subjects

Yajing Wang

Publications and source records attributed to Yajing Wang.

10 recordsLinked to original sources

arXiv2026

When LLM Rationales Become User-Facing: Effects on Trust Perception, Decision-Making, and Gaze Behaviors

Large language models (LLMs) increasingly show step-by-step reasoning rationales alongside their answers, turning reasoning from an internal model capability into a user-facing interface feature. Yet it is unclear whether such rationales help users judge when trust is warranted or merely persuade through fluent reasoning. We address this gap through the lens of auditable trust calibration: user-facing rationales should help people inspect whether an answer is warranted by evidence. We test this framing in factual verification through two linked studies. Study 1, an online experiment (N=68), manipulated rationale presentation format (instant, delayed, on demand), rationale correctness (correct, incorrect), and certainty framing (none, certain, uncertain). Study 2, a controlled eye-tracking study (N=54), examined how no-, correct-, and incorrect-rationale conditions were associated with users' trust, decision-making, and eye-movement patterns. Study 1 showed no reliable presentation-format effects; instead, rationale correctness and certainty framing influenced the trust in the information, trust in the LLM system, and decision confidence. In Study 2, incorrect rationales drew more attention to the supporting evidence and larger pupil diameter while the rationale was viewed, consistent with greater cognitive effort. Incorrect rationales also lowered trust in LLM system relative to showing no rationale, whereas the no-rationale difference was weaker for trust in information. A post-hoc predictive modeling analysis of gaze data from Study 2 further showed that gaze features carried predictive signal for trust- and decision-related user states. This work challenges the assumption that more reasoning is always better and supports rationale designs that are selective, linked to evidence, calibrated in how they express certainty, and easier to verify.

cs.HC↗

arXiv2026

ActiveScope: Actively Seeking and Correcting Perception for MLLMs

Multimodal Large Language Models (MLLMs) have demonstrated impressive vision-language understanding, yet still struggle with fine-grained perception in high-resolution images. While existing training-free methods typically rely on attention-based localization or coarse-to-fine search, they are often misled by distractors and fail to locate multiple targets. Our investigation attributes these failures to Contextual Dominance, where salient distractors overwhelm target attention and cause inaccurate localization, and Semantic Bias, where global semantics cause the model to fixate on the most salient concept, resulting in incomplete localization in multi-object scenarios. Built on these insights, we propose ActiveScope, a training-free framework that enhances MLLMs by actively seeking and correcting perception. ActiveScope features two modules. The Semantic Anchor Localization (SAL) utilizes fine-grained semantic anchors to independently localize key targets, thereby mitigating semantic bias. The Interference-Suppressed Refinement (ISR) refines localization by suppressing attention on salient distractions to overcome contextual dominance. Extensive experiments on high-resolution image understanding benchmarks demonstrate that ActiveScope outperforms existing training-free methods (e.g., 96.34 percent accuracy on $V^{*}$ Bench), validating the superiority of the active search and self-correction paradigm. Our code is available at https://github.com/jasmine-ww/ActiveScope.

cs.CV↗

arXiv2026

Eyes Can't Always Tell: Fusing Eye Tracking and User Priors for User Modeling under AI Advice Conditions

Modeling users' cognitive states (e.g., cognitive load and decision confidence) is essential for building adaptive AI in high-stakes decision-making. While eye tracking provides non-invasive behavioral signals correlated with cognitive effort, prior work has not systematically examined how AI assistance contexts, specifically varying advice reliability and user heterogeneity, can alter the mapping between gaze signals and cognitive states. We conducted a within-subject lab eye-tracking study (N=54) on factual verification tasks under three conditions: No-AI, Correct-AI advice, and Incorrect-AI advice. We analyze condition-dependent changes in self-reports and eye-tracking patterns and evaluate the robustness of eye-tracking-based user modeling. Results show that AI advice increases decision confidence compared to No-AI, while Correct-AI is associated with lower perceived cognitive load and more efficient gaze behavior. Crucially, predictive modeling is context-sensitive: the relationship between eye-tracking signals and cognitive states shifts across AI conditions. Finally, fusing eye-tracking features with user priors (demographics, AI literacy/experience, and propensity to trust technology) improves cross-participant generalization. These findings support condition-aware and personalized user modeling for cognitively aligned adaptive AI systems.

cs.HC↗

arXiv2025

Lonely Individuals Show Distinct Patterns of Social Media Engagement

Loneliness has reached epidemic proportions globally, posing serious risks to mental and physical health. As social media platforms increasingly mediate social interaction, understanding their relationship with loneliness has become urgent. While survey-based research has examined social media use and loneliness, findings remain mixed, and little is known about when and how often people engage with social media, or about whether different types of platforms are differently associated with loneliness. Web trace data now enable objective examination of these behavioral dimensions. We asked whether objectively measured patterns of social media engagement differ between lonely and non-lonely individuals across devices and platform types. Analyzing six months of web trace data combined with repeated surveys ($N=589$ mobile users; $N=851$ desktop users), we found that greater social media use was associated with higher loneliness across both devices, with this relationship specific to social media rather than other online activities. On desktop, lonely individuals exhibited shorter sessions but more frequent daily engagement. Lonely individuals spent more time on visual-sharing ($g = -0.47$), messaging ($g = -0.36$), and networking-oriented platforms on mobile. These findings demonstrate how longitudinal web trace data can reveal behavioral patterns associated with loneliness, and more broadly illustrate the potential of digital traces for studying other psychological states. Beyond research, the results inform the responsible design of digital interventions and platform features that better support psychological well-being across different technological contexts.

cs.HC↗

arXiv2025

Self-Guided Function Calling in Large Language Models via Stepwise Experience Recall

Function calling enables large language models (LLMs) to interact with external systems by leveraging tools and APIs. When faced with multi-step tool usage, LLMs still struggle with tool selection, parameter generation, and tool-chain planning. Existing methods typically rely on manually designing task-specific demonstrations, or retrieving from a curated library. These approaches demand substantial expert effort and prompt engineering becomes increasingly complex and inefficient as tool diversity and task difficulty scale. To address these challenges, we propose a self-guided method, Stepwise Experience Recall (SEER), which performs fine-grained, stepwise retrieval from a continually updated experience pool. Instead of relying on static or manually curated library, SEER incrementally augments the experience pool with past successful trajectories, enabling continuous expansion of the pool and improved model performance over time. Evaluated on the ToolQA benchmark, SEER achieves an average improvement of 6.1% on easy and 4.7% on hard questions. We further test SEER on $\tau$-bench, which includes two real-world domains. Powered by Qwen2.5-7B and Qwen2.5-72B models, SEER demonstrates substantial accuracy gains of 7.44% and 23.38%, respectively.

cs.CL↗

arXiv2025

Advancing Tumor Budding Detection with Fourier Ptychography Microscopy

Background: Tumour budding is an independent predictor of metastasis and prognosis in colorectal cancer and is a vital part of the pathology specification report. In a conventional pathological section observation process, pathologists have to repeatedly switch from 10x objective to 20x objective several times to localize and image the target region. Besides the switching operations, repeated manual or electro-mechanical focusing is also very time-consuming, affecting the total time for pathological diagnosis. In addition, It is usually necessary to remove the manually marked symbols on the stained pathology slides used for classification and management before observation. Methods: In this paper, we utilize Fourier ptychographic microscopy (FPM) in the pathological diagnosis process to realize large space-bandwidth product imaging, quantitative phase imaging, and digital refocusing in the observation process without any mechanical operations, which can therefore simplify the above-mentioned cumbersome diagnostic processes. We first verify the effectiveness and efficiency of the proposed method with several typical pathological sections. Then, instead of manually erasing, we also prove that FP framework can digitally remove the artificial markers with its digital refocusing ability. Results: At last, we demonstrated pathologists can achieve 100% diagnostic accuracy with FPM imaging results. Conclusions: The proposed method can greatly simplify the process of pathological diagnosis, and the related addon hardware system does not require expensive components, which makes it have great potential for promotion in the field of pathological diagnosis.

physics.med-ph↗

arXiv2024

Eliciting Causal Abilities in Large Language Models for Reasoning Tasks

Prompt optimization automatically refines prompting expressions, unlocking the full potential of LLMs in downstream tasks. However, current prompt optimization methods are costly to train and lack sufficient interpretability. This paper proposes enhancing LLMs' reasoning performance by eliciting their causal inference ability from prompting instructions to correct answers. Specifically, we introduce the Self-Causal Instruction Enhancement (SCIE) method, which enables LLMs to generate high-quality, low-quantity observational data, then estimates the causal effect based on these data, and ultimately generates instructions with the optimized causal effect. In SCIE, the instructions are treated as the treatment, and textual features are used to process natural language, establishing causal relationships through treatments between instructions and downstream tasks. Additionally, we propose applying Object-Relational (OR) principles, where the uncovered causal relationships are treated as the inheritable class across task objects, ensuring low-cost reusability. Extensive experiments demonstrate that our method effectively generates instructions that enhance reasoning performance with reduced training cost of prompts, leveraging interpretable textual features to provide actionable insights.

cs.CL↗

arXiv2024

Enhance Multi-domain Sentiment Analysis of Review Texts through Prompting Strategies

Large Language Models (LLMs) have made significant strides in both scientific research and practical applications. Existing studies have demonstrated the state-of-the-art (SOTA) performance of LLMs in various natural language processing tasks. However, the question of how to further enhance LLMs' performance in specific task using prompting strategies remains a pivotal concern. This paper explores the enhancement of LLMs' performance in sentiment analysis through the application of prompting strategies. We formulate the process of prompting for sentiment analysis tasks and introduce two novel strategies tailored for sentiment analysis: RolePlaying (RP) prompting and Chain-of-thought (CoT) prompting. Specifically, we also propose the RP-CoT prompting strategy which is a combination of RP prompting and CoT prompting. We conduct comparative experiments on three distinct domain datasets to evaluate the effectiveness of the proposed sentiment analysis strategies. The results demonstrate that the adoption of the proposed prompting strategies leads to a increasing enhancement in sentiment analysis accuracy. Further, the CoT prompting strategy exhibits a notable impact on implicit sentiment analysis, with the RP-CoT prompting strategy delivering the most superior performance among all strategies.

cs.CL↗

arXiv2020

A Hierarchical Collision Avoidance Architecture for Multiple Fixed-Wing UAVs in an Integrated Airspace

This paper studies the collision avoidance problem for autonomous multiple fixedwing UAVs in the complex integrated airspace. By studying and combining the online path planning method, the distributed model predictive control algorithm, and the geometric reactive control approach, a three-layered collision avoidance system integrating conflict detection and resolution procedures is developed for multiple fixed-wing UAVs modeled by unicycle kinematics subject to input constraints. The effectiveness of the proposed methodology is evaluated and validated via test results of comparative simulations under both deterministic and probabilistic sensing conditions.

cs.RO↗

arXiv2019

Flexible Demand Resource Pricing Scheme: A Stochastic Benefit-Sharing Approach

With the rapidly increased penetration of renewable generations, incentive-based demand side management (DSM) shows great value on alleviating the uncertainty and providing flexibility for microgrid. However, how to price those demand resources becomes one of the most significant challenges for promoting incentive-based DSM under microgrid environments. In this paper, a flexible demand resource pricing scheme is proposed. Instead of using the utility function of end users like most existing literatures, the economic benefit of flexible demand resources is evaluated by the operation performance enhancement of microgrid and correspondingly the resource is priced based on a benefit sharing approach. An iteration-based chance-constrained method is established to calculate the economic benefit and shared compensation for demand resource providers. Meanwhile, the financial risks for the microgrid operator due to uncertain factors are mitigated by the chance-constrained criterion. The proposed scheme is examined by an experimental microgrid to illustrate its effectiveness.

math.OC↗