arXiv Science⌕ Search

arXiv subjects

Lingyao Li

Publications and source records attributed to Lingyao Li.

At least 73 records · Page 4Linked to original sources

Towards Trustworthy AI: Characterizing User-Reported Risks across LLMs "In the Wild"

While Large Language Models (LLMs) are rapidly integrating into daily life, research on their risks often remains lab-based and disconnected from the problems users encounter "in the wild." While recent HCI research has begun to explore these user-facing risks, it typically concentrates on a singular LLM chatbot like ChatGPT or an isolated risk like privacy. To gain a holistic understanding of multi-risk across LLM chatbots, we analyze online discussions on Reddit around seven major LLM chatbots through the U.S. NIST's AI Risk Management Framework. We find that user-reported risks are unevenly distributed and platform-specific. While "Valid and Reliable" risk is the most frequently mentioned, each product also exhibits a unique "risk fingerprint;" for instance, user discussions associate GPT more with "Safe" and "Fair" issues, Gemini with "Privacy," and Claude with "Secure and Resilient" risks. Furthermore, the nature of these risks differs by their prevalence: less frequent risks like "Explainability" and "Privacy" manifest as nuanced user trade-offs, more common ones like "Fairness" are experienced as direct personal harms. Our findings reveal gaps between risks reported by system-centered studies and by users, highlighting the need for user-centered approaches that support users in their daily use of LLM chatbots.

cs.CY↗

NPHardEval4V: Dynamic Evaluation of Large Vision-Language Models with Effects of Vision

Large Vision-Language Models (LVLMs) have demonstrated impressive capabilities in multimodal understanding, yet their reasoning abilities remain underexplored. Existing benchmarks tend to focus on perception or text-based comprehension, offering limited insight into how well these models perform on structured, logic-driven tasks that require both visual and linguistic reasoning. To address this gap, we introduce NPHardEval4V, a multimodal benchmark suite grounded in four classical NP-hard problems: Knapsack, Set Cover, Traveling Salesperson, and Vertex Cover. Each task is presented through a combination of structured visual layouts and textual prompts, designed to assess the ability of LVLMs to perform combinatorial reasoning under visual-linguistic constraints. We evaluate a set of advanced open-source and closed-source vision-language models under a unified prompting and problem representation framework. This enables fair comparison across models and task types, while isolating key variables affecting performance. Our results show that while these models perform reasonably well on perception-based inputs, they struggle with global optimization, abstraction, and constraint satisfaction. No single model demonstrates consistent reasoning capability across all problem types, and common failure patterns reveal fundamental limitations in current architectures. By leveraging the structure and complexity of NP-hard problems, NPHardEval4V provides a scalable, interpretable, and challenging testbed for diagnosing reasoning behaviors in LVLMs. We hope this benchmark can support the community in building more robust, inference-capable multimodal systems. The benchmark dataset and code are available at https://github.com/lizhouf/NPHardEval4.

cs.CL↗

RAG for Geoscience: What We Expect, Gaps and Opportunities

Retrieval-Augmented Generation (RAG) enhances language models by combining retrieval with generation. However, its current workflow remains largely text-centric, limiting its applicability in geoscience. Many geoscientific tasks are inherently evidence-hungry. Typical examples involve imputing missing observations using analog scenes, retrieving equations and parameters to calibrate models, geolocating field photos based on visual cues, or surfacing historical case studies to support policy analyses. A simple ``retrieve-then-generate'' pipeline is insufficient for these needs. We envision Geo-RAG, a next-generation paradigm that reimagines RAG as a modular retrieve $\rightarrow$ reason $\rightarrow$ generate $\rightarrow$ verify loop. Geo-RAG supports four core capabilities: (i) retrieval of multi-modal Earth data; (ii) reasoning under physical and domain constraints; (iii) generation of science-grade artifacts; and (iv) verification of generated hypotheses against numerical models, ground measurements, and expert assessments. This shift opens new opportunities for more trustworthy and transparent geoscience workflows.

cs.ET↗

From Promising Capability to Pervasive Bias: Assessing Large Language Models for Emergency Department Triage

Large Language Models (LLMs) have shown promise in clinical decision support, yet their application to triage remains underexplored. We systematically investigate the capabilities of LLMs in emergency department triage through two key dimensions: (1) robustness to distribution shifts and missing data, and (2) counterfactual analysis of intersectional biases across sex and race. We assess multiple LLM-based approaches, ranging from continued pre-training to in-context learning, as well as machine learning approaches. Our results indicate that LLMs exhibit superior robustness, and we investigate the key factors contributing to the promising LLM-based approaches. Furthermore, in this setting, we identify gaps in LLM preferences that emerge in particular intersections of sex and race. LLMs generally exhibit sex-based differences, but they are most pronounced in certain racial groups. These findings suggest that LLMs encode demographic preferences that may emerge in specific clinical contexts or particular combinations of characteristics.

cs.AI↗

DynamiCare: A Dynamic Multi-Agent Framework for Interactive and Open-Ended Medical Decision-Making

The rise of Large Language Models (LLMs) has enabled the development of specialized AI agents with domain-specific reasoning and interaction capabilities, particularly in healthcare. While recent frameworks simulate medical decision-making, they largely focus on single-turn tasks where a doctor agent receives full case information upfront -- diverging from the real-world diagnostic process, which is inherently uncertain, interactive, and iterative. In this paper, we introduce MIMIC-Patient, a structured dataset built from the MIMIC-III electronic health records (EHRs), designed to support dynamic, patient-level simulations. Building on this, we propose DynamiCare, a novel dynamic multi-agent framework that models clinical diagnosis as a multi-round, interactive loop, where a team of specialist agents iteratively queries the patient system, integrates new information, and dynamically adapts its composition and strategy. We demonstrate the feasibility and effectiveness of DynamiCare through extensive experiments, establishing the first benchmark for dynamic clinical decision-making with LLM-powered agents.

cs.AI↗

Integrating LLMs and Digital Twins for Adaptive Multi-Robot Task Allocation in Construction

Multi-robot systems are emerging as a promising solution to the growing demand for productivity, safety, and adaptability across industrial sectors. However, effectively coordinating multiple robots in dynamic and uncertain environments, such as construction sites, remains a challenge, particularly due to unpredictable factors like material delays, unexpected site conditions, and weather-induced disruptions. To address these challenges, this study proposes an adaptive task allocation framework that strategically leverages the synergistic potential of Digital Twins, Integer Programming (IP), and Large Language Models (LLMs). The multi-robot task allocation problem is formally defined and solved using an IP model that accounts for task dependencies, robot heterogeneity, scheduling constraints, and re-planning requirements. A mechanism for narrative-driven schedule adaptation is introduced, in which unstructured natural language inputs are interpreted by an LLM, and optimization constraints are autonomously updated, enabling human-in-the-loop flexibility without manual coding. A digital twin-based system has been developed to enable real-time synchronization between physical operations and their digital representations. This closed-loop feedback framework ensures that the system remains dynamic and responsive to ongoing changes on site. A case study demonstrates both the computational efficiency of the optimization algorithm and the reasoning performance of several LLMs, with top-performing models achieving over 97% accuracy in constraint and parameter extraction. The results confirm the practicality, adaptability, and cross-domain applicability of the proposed methods.

cs.RO↗

A Multimodal, Multilingual, and Multidimensional Pipeline for Fine-grained Crowdsourcing Earthquake Damage Evaluation

Rapid, fine-grained disaster damage assessment is essential for effective emergency response, yet remains challenging due to limited ground sensors and delays in official reporting. Social media provides a rich, real-time source of human-centric observations, but its multimodal and unstructured nature presents challenges for traditional analytical methods. In this study, we propose a structured Multimodal, Multilingual, and Multidimensional (3M) pipeline that leverages multimodal large language models (MLLMs) to assess disaster impacts. We evaluate three foundation models across two major earthquake events using both macro- and micro-level analyses. Results show that MLLMs effectively integrate image-text signals and demonstrate a strong correlation with ground-truth seismic data. However, performance varies with language, epicentral distance, and input modality. This work highlights the potential of MLLMs for disaster assessment and provides a foundation for future research in applying MLLMs to real-time crisis contexts. The code and data are released at: https://github.com/missa7481/EMNLP25_earthquake

cs.CL↗

LLMs as World Models: Data-Driven and Human-Centered Pre-Event Simulation for Disaster Impact Assessment

Efficient simulation is essential for enhancing proactive preparedness for sudden-onset disasters such as earthquakes. Recent advancements in large language models (LLMs) as world models show promise in simulating complex scenarios. This study examines multiple LLMs to proactively estimate perceived earthquake impacts. Leveraging multimodal datasets including geospatial, socioeconomic, building, and street-level imagery data, our framework generates Modified Mercalli Intensity (MMI) predictions at zip code and county scales. Evaluations on the 2014 Napa and 2019 Ridgecrest earthquakes using USGS ''Did You Feel It? (DYFI)'' reports demonstrate significant alignment, as evidenced by a high correlation of 0.88 and a low RMSE of 0.77 as compared to real reports at the zip code level. Techniques such as RAG and ICL can improve simulation performance, while visual inputs notably enhance accuracy compared to structured numerical data alone. These findings show the promise of LLMs in simulating disaster impacts that can help strengthen pre-event planning.

cs.CY↗

Invisible Prompts, Visible Threats: Malicious Font Injection in External Resources for Large Language Models

Large Language Models (LLMs) are increasingly equipped with capabilities of real-time web search and integrated with protocols like Model Context Protocol (MCP). This extension could introduce new security vulnerabilities. We present a systematic investigation of LLM vulnerabilities to hidden adversarial prompts through malicious font injection in external resources like webpages, where attackers manipulate code-to-glyph mapping to inject deceptive content which are invisible to users. We evaluate two critical attack scenarios: (1) "malicious content relay" and (2) "sensitive data leakage" through MCP-enabled tools. Our experiments reveal that indirect prompts with injected malicious font can bypass LLM safety mechanisms through external resources, achieving varying success rates based on data sensitivity and prompt design. Our research underscores the urgent need for enhanced security measures in LLM deployments when processing external content.

cs.CR↗

Know the Ropes: A Heuristic Strategy for LLM-based Multi-Agent System Design

Single-agent LLMs hit hard limits--finite context, role overload, and brittle domain transfer. Conventional multi-agent fixes soften those edges yet expose fresh pains: ill-posed decompositions, fuzzy contracts, and verification overhead that blunts the gains. We therefore present Know-The-Ropes (KtR), a framework that converts domain priors into an algorithmic blueprint hierarchy, in which tasks are recursively split into typed, controller-mediated subtasks, each solved zero-shot or with the lightest viable boost (e.g., chain-of-thought, micro-tune, self-check). Grounded in the No-Free-Lunch theorem, KtR trades the chase for a universal prompt for disciplined decomposition. On the Knapsack problem (3-8 items), three GPT-4o-mini agents raise accuracy from 3% zero-shot to 95% on size-5 instances after patching a single bottleneck agent. On the tougher Task-Assignment problem (6-15 jobs), a six-agent o3-mini blueprint hits 100% up to size 10 and 84% on sizes 13-15, versus 11% zero-shot. Algorithm-aware decomposition plus targeted augmentation thus turns modest models into reliable collaborators--no ever-larger monoliths required.

cs.AI↗

AttackEval: How to Evaluate the Effectiveness of Jailbreak Attacking on Large Language Models

Jailbreak attacks represent one of the most sophisticated threats to the security of large language models (LLMs). To deal with such risks, we introduce an innovative framework that can help evaluate the effectiveness of jailbreak attacks on LLMs. Unlike traditional binary evaluations focusing solely on the robustness of LLMs, our method assesses the attacking prompts' effectiveness. We present two distinct evaluation frameworks: a coarse-grained evaluation and a fine-grained evaluation. Each framework uses a scoring range from 0 to 1, offering unique perspectives and allowing for the assessment of attack effectiveness in different scenarios. Additionally, we develop a comprehensive ground truth dataset specifically tailored for jailbreak prompts. This dataset is a crucial benchmark for our current study and provides a foundational resource for future research. By comparing with traditional evaluation methods, our study shows that the current results align with baseline metrics while offering a more nuanced and fine-grained assessment. It also helps identify potentially harmful attack prompts that might appear harmless in traditional evaluations. Overall, our work establishes a solid foundation for assessing a broader range of attack prompts in prompt injection.

cs.CL↗

Instilling Doubts About Truth: Measuring the Impact of Tucker Carlson's Interview with Vladimir Putin Using Machine Learning and Natural Language Processing

On February 7, 2024, Russian President Vladimir Putin gave a two-hour interview with conservative political commentator, Tucker Carlson. This study investigated the impact of the Carlson- Putin interview on the US X audience. We proposed a framework of social media impact using machine learning (ML) and natural language processing (NLP) by measuring changes in audience, structure, and content. Triangulation methods were used to validate the process and results. The interview had a considerable impact among segments of the American public: 1) the reach and engagement of far-right influencers increased after the interview, suggesting Kremlin narratives gained traction within these circles, 2) the communication structure became more vulnerable to disinformation spread after the interview, and 3) the public discourse changed from support for Ukraine funding to conversations about Putin, Russia, and the issue of "truth" or the veracity of Putin's claims. This research contributes to methods development for social media studies and aids scholars in analyzing how public opinion shapes policy debates. The Carlson-Putin interview sparked a broader discussion about truth-telling. Far from being muted, the broad impact of the interview appears considerable and poses challenges for foreign affairs leaders who depend on public support and buy-in when formulating national policy.

cs.SI↗

ADO: Automatic Data Optimization for Inputs in LLM Prompts

This study explores a novel approach to enhance the performance of Large Language Models (LLMs) through the optimization of input data within prompts. While previous research has primarily focused on refining instruction components and augmenting input data with in-context examples, our work investigates the potential benefits of optimizing the input data itself. We introduce a two-pronged strategy for input data optimization: content engineering and structural reformulation. Content engineering involves imputing missing values, removing irrelevant attributes, and enriching profiles by generating additional information inferred from existing attributes. Subsequent to content engineering, structural reformulation is applied to optimize the presentation of the modified content to LLMs, given their sensitivity to input format. Our findings suggest that these optimizations can significantly improve the performance of LLMs in various tasks, offering a promising avenue for future research in prompt engineering. The source code is available at https://anonymous.4open.science/r/ADO-6BC5/

cs.LG↗

Sympathy over Polarization: A Computational Discourse Analysis of Social Media Posts about the July 2024 Trump Assassination Attempt

On July 13, 2024, at the Trump rally in Pennsylvania, someone attempted to assassinate Republican Presidential Candidate Donald Trump. This attempt sparked a large-scale discussion on social media. We collected posts from X (formerly known as Twitter) one week before and after the assassination attempt and aimed to model the short-term effects of such a ``shock'' on public opinions and discussion topics. Specifically, our study addresses three key questions: first, we investigate how public sentiment toward Donald Trump shifts over time and across regions (RQ1) and examine whether the assassination attempt itself significantly affects public attitudes, independent of the existing political alignments (RQ2). Finally, we explore the major themes in online conversations before and after the crisis, illustrating how discussion topics evolved in response to this politically charged event (RQ3). By integrating large language model-based sentiment analysis, difference-in-differences modeling, and topic modeling techniques, we find that following the attempt the public response was broadly sympathetic to Trump rather than polarizing, despite baseline ideological and regional disparities.

cs.SI↗

Academic collaboration on large language model studies increases overall but varies across disciplines

Interdisciplinary collaboration is crucial for addressing complex scientific challenges. Recent advancements in large language models (LLMs) have shown significant potential in benefiting researchers across various fields. To explore their potential for interdisciplinary collaboration, we collect and analyze data from OpenAlex, an open-source academic database. Our dataset comprises 59,293 LLM-related papers, along with 70,945 machine learning (ML) papers and 73,110 papers from non-LLM/ML fields as control groups. We first employ Shannon Entropy to assess the diversity of collaboration. Our results reveal that many fields have exhibited a more significant increasing trend following the release of ChatGPT as compared to the control groups. In particular, Computer Science and Social Science display a consistent increase in both institution and department entropy. Other fields such as Decision Science, Psychology, and Health Professions have shown minor to significant increases. Our difference-in-difference analysis also indicates that the release of ChatGPT leads to a statistically significant increase in collaboration in several fields, such as Computer Science and Social Science. In addition, we analyze the author networks and find that Computer Science, Medicine, and other Computer Science-related departments are the most prominent. Regarding authors' institutions, our analysis reveals that entities such as Stanford University, Harvard University, and University College London are key players, either dominating centrality or playing crucial roles in connecting research networks. Overall, this study provides valuable information on the current landscape and evolving dynamics of collaboration networks in LLM research. It also suggests potential areas for fostering more diverse collaborations and highlights the need for continued research on the impact of LLMs on scientific practices.

cs.SI↗

Assessing Response Disparities in California Wildland-Urban-Interface (WUI) Cities Using the Compartmental Model

The increasing frequency and severity of wildfires pose significant risks to communities, infrastructure, and the environment, especially in Wildland-Urban Interface (WUI) areas. Effective disaster management requires understanding how the public perceives and responds to wildfire threats in real-time. This study uses social media data to assess public responses and explores how these responses are linked to city-level community characteristics. Specifically, we leveraged a transformer-based topic modeling technique called BERTopic to identify wildfire response-related topics and then utilized the Susceptible-Infectious-Recovered (SIR) model to compute two key metrics associated with wildfire responses - awareness and resilience indicators. Additionally, we used GIS-based spatial analysis to map wildfire locations along with four groups of city-level factors (racial/ethnic, socioeconomic, demographic, and wildfire-specific). Our findings reveal significant geographic and socio-spatial differences in public responses. Southern California cities with larger Hispanic populations demonstrate higher wildfire awareness and resilience. In contrast, urbanized regions in Central and Northern California exhibit lower awareness levels. Furthermore, resilience is negatively correlated with unemployment rates, particularly in southern regions where higher unemployment aligns with reduced resilience. These findings highlight the need for targeted and equitable wildfire management strategies to improve the adaptive capacity of WUI communities.

cs.SI↗

Game-theoretic LLM: Agent Workflow for Negotiation Games

This paper investigates the rationality of large language models (LLMs) in strategic decision-making contexts, specifically within the framework of game theory. We evaluate several state-of-the-art LLMs across a spectrum of complete-information and incomplete-information games. Our findings reveal that LLMs frequently deviate from rational strategies, particularly as the complexity of the game increases with larger payoff matrices or deeper sequential trees. To address these limitations, we design multiple game-theoretic workflows that guide the reasoning and decision-making processes of LLMs. These workflows aim to enhance the models' ability to compute Nash Equilibria and make rational choices, even under conditions of uncertainty and incomplete information. Experimental results demonstrate that the adoption of these workflows significantly improves the rationality and robustness of LLMs in game-theoretic tasks. Specifically, with the workflow, LLMs exhibit marked improvements in identifying optimal strategies, achieving near-optimal allocations in negotiation scenarios, and reducing susceptibility to exploitation during negotiations. Furthermore, we explore the meta-strategic considerations of whether it is rational for agents to adopt such workflows, recognizing that the decision to use or forgo the workflow constitutes a game-theoretic issue in itself. Our research contributes to a deeper understanding of LLMs' decision-making capabilities in strategic contexts and provides insights into enhancing their rationality through structured workflows. The findings have implications for the development of more robust and strategically sound AI agents capable of navigating complex interactive environments. Code and data supporting this study are available at \url{https://github.com/Wenyueh/game_theory}.

cs.AI↗

Characterizing Online Toxicity During the 2022 Mpox Outbreak: A Computational Analysis of Topical and Network Dynamics

Background: Online toxicity, encompassing behaviors such as harassment, bullying, hate speech, and the dissemination of misinformation, has become a pressing social concern in the digital age. The 2022 Mpox outbreak, initially termed "Monkeypox" but subsequently renamed to mitigate associated stigmas and societal concerns, serves as a poignant backdrop to this issue. Objective: In this research, we undertake a comprehensive analysis of the toxic online discourse surrounding the 2022 Mpox outbreak. Our objective is to dissect its origins, characterize its nature and content, trace its dissemination patterns, and assess its broader societal implications, with the goal of providing insights that can inform strategies to mitigate such toxicity in future crises. Methods: We collected more than 1.6 million unique tweets and analyzed them from five dimensions, including context, extent, content, speaker, and intent. Utilizing BERT-based topic modeling and social network community clustering, we delineated the toxic dynamics on Twitter. Results: We identified five high-level topic categories in the toxic online discourse on Twitter, including disease (46.6%), health policy and healthcare (19.3%), homophobia (23.9%), politics (6.0%), and racism (4.1%). Through the toxicity diffusion networks of mentions, retweets, and the top users, we found that retweets of toxic content were widespread, while influential users rarely engaged with or countered this toxicity through retweets. Conclusions: By tracking topical dynamics, we can track the changing popularity of toxic content online, providing a better understanding of societal challenges. Network dynamics spotlight key social media influencers and their intents, indicating that addressing these central figures in toxic discourse can enhance crisis communication and inform policy-making.

cs.SI↗