论文检索

Yilun Liu, Chunguang Zhao, Xinhua Yang, Hongyong Zeng, Shimin Tao, Weibin Meng, Minggui He, Yan Yu, Hongxia Ma, Li Zhang 等

Despite doubts on data quality, instruction synthesis has been widely applied into instruction tuning (IT) of LLMs as an economic and rapid alternative. Recent endeavors focus on improving data quality for synthesized instruction pairs in English and have facilitated IT of English-centric LLMs. However, data quality issues in multilingual synthesized instruction pairs are even more severe, since the common synthesizing practice is to translate English synthesized data into other languages using machine translation (MT). Besides the known content errors in these English synthesized data, multilingual synthesized instruction data are further exposed to defects introduced by MT and face insufficient localization of the target languages, leading to cultural inequality in trained LLMs. In this paper, we propose MIDB, a Multilingual Instruction Data Booster to automatically address the quality issues in multilingual synthesized data. MIDB is trained on around 36.8k revision examples across 16 languages by human linguistic experts, thereby can boost the low-quality data by addressing content errors and MT defects, and improving localization in these synthesized data. Both automatic and human evaluation indicate that not only MIDB steadily improved instruction data quality in 16 languages, but also the instruction-following and cultural-understanding abilities of multilingual LLMs fine-tuned on MIDB-boosted data were significantly enhanced, suggesting an improved linguistic and cultural equality.

Xin Liu, Yuanhang Yu, Peng Zhu, Dawei Cheng, Changjun Jiang

Illegal related-party transactions (RPT) are federal felonies that pose a severe threat to the stability and integrity of modern financial systems. The increasing frequency of RPTs forms complex and dynamic networks. Existing temporal graph learning methods tend to treat entities as functionally homogeneous, ignoring the diverse and evolving structural roles of nodes. Role-based embedding methods model global structure by bridging same-role nodes, but their reliance on a unified mechanism for aggregation and evolution means they fail to distinguish the underlying logic of distinct interactions governed by structural roles. The limitations motivate us to develop a customized role-based strategy. It can also adapt to evolving RPT dynamics, thereby forming a continuous regulatory process to combat illegal activities. In this paper, we propose an innovative Role Perceptual Augmented Temporal Graph Network (RPATGN) for proactive RPT detection. We analyze the structural roles of nodes and employ a role-based message passing mechanism that adapts its aggregation strategy based on the roles of interacting nodes. We integrate a variational graph recurrent neural network, enhanced by temporal contextual attention, to explicitly model the dynamics of the roles and the overall network evolution. Extensive experiments on real-world financial datasets demonstrate the effectiveness of our approach for RPT detection. It holds practical significance for fostering robust financial systems and promoting healthy, transparent financial markets.

Liu Liu, Alexandra Schild, Marco Cipriano, Fatimeh Al Ghannam, Freya Tan, Gerard De Melo, Andres Sevtsuk

Understanding group-level social interactions in public spaces is crucial for urban planning, informing the design of socially vibrant and inclusive environments. Detecting such interactions from images involves interpreting subtle visual cues such as relations, proximity and co-movement – semantically complex signals that go beyond traditional object detection. To address this challenge, we introduce a social group region detection task, which requires inferring and spatially grounding visual regions defined by abstract interpersonal relations. We propose MINGLE (Modeling INterpersonal Group-Level Engagement), a modular three-stage pipeline that integrates: (1) off-the-shelf human detection and depth estimation, (2) VLM-based reasoning to classify pairwise social affiliation, and (3) a lightweight spatial aggregation algorithm to localize socially connected groups. To support this task and encourage future research, we present a new dataset of 100K urban street-view images annotated with bounding boxes and labels for both individuals and socially interacting groups. The annotations combine human-created labels and outputs from the MINGLE pipeline, ensuring semantic richness and broad coverage of real world scenarios.

Jiangfan Liu, Yongkang Guo, Fangzhi Zhong, Tianyuan Zhang, Zonglei Jing, Siyuan Liang, Jiakai Wang, Mingchuan Zhang, Aishan Liu, Xianglong Liu

The generation of safety-critical scenarios in simulation has become increasingly crucial for safety evaluation in autonomous vehicles (AV) prior to road deployment in society. However, current approaches largely rely on predefined threat patterns or rule-based strategies, which limit their ability to expose diverse and unforeseen failure modes. To overcome these, we propose ScenGE, a framework that can generate plentiful safety-critical scenarios by reasoning novel adversarial cases and then amplifying them with complex traffic flows. Given a simple prompt of a benign scene, it first performs Meta-Scenario Generation, where a large language model (LLM), grounded in structured driving knowledge (e.g., traffic regulations, real-world accident records), infers an adversarial agent whose behavior poses a threat that is both plausible and deliberately challenging. This meta-scenario is then specified in executable code for precise in-simulator control. Subsequently, Complex Scenario Evolution uses background vehicles to amplify the core threat introduced by Meta-Scenario. It builds an adversarial collaborator graph to identify key agent trajectories for optimization. These perturbations are designed to simultaneously reduce the ego vehicle's maneuvering space and create critical occlusions. Extensive experiments conducted on multiple reinforcement learning (RL) based AV models show that ScenGE uncovers more severe collision cases (+31.96%) on average than SoTA baselines. Additionally, our ScenGE can be applied to large model based AV systems and deployed on different simulators; we further observe that adversarial training on our scenarios improves the model robustness. We hope our paper can build up a critical step towards building public trust and ensuring their safe deployment.

Anush Lingamoorthy, Abhishek Kumar Mishra, Olumuyiwa Oni, Jacob S Brenner, Nagarajan Kandasamy, Amanda Watson

Opioid overdose is a growing global health crisis that claims more than 120,000 lives annually, of which more than half use opioids alone, without access to bystander intervention. Fatal overdose events are marked by motionlessness, respiratory depression, and hypoxemia, yet current wearable systems often rely on a single biomarker, limiting detection speed and accuracy. We present HypoxSpike, a novel ternary spiking neural network designed for real-time, multi-biomarker overdose detection for low-power neuromorphic hardware, optimized for integration into shoulder-based wearables. HypoxSpike combines motion, respiration, and oxygen saturation signals, while accounting for skin tone and body physiology, thus addressing known racial bias in pulse oximetry. Our research leverages an open-source shoulder-worn dataset from 19 patients experiencing sleep apnea, exploiting the shared physiological mechanisms underlying apnea and opioid overdose. This allows a direct comparison of our model with existing overdose detection approaches. HypoxSpike classifies three stages of hypoxemia with an average accuracy of 94%, outperforming state-of-the-art shoulder-based hypoxemia estimation while reducing false positive alert rates by 23.5%. By minimizing false positives, HypoxSpike supports accurate and power-efficient overdose detection, improving trust and usability for high-risk populations often overlooked by conventional systems.

Chenxi Lin, Weikang Yuan, Zhuoren Jiang, Biao Huang, Ruitao Zhang, Jianan Ge, Yueqian Xu, Jianxing Yu

Understanding human attitudes, preferences, and behaviors through social surveys is essential for academic research and policymaking. Yet traditional surveys face persistent challenges, including fixed-question formats, high costs, limited adaptability, and difficulties ensuring cross-cultural equivalence. While recent studies explore large language models (LLMs) to simulate survey responses, most are limited to structured questions, overlook the entire survey process, and risks under-representing marginalized groups due to training data biases. We introduce AlignSurvey, the first benchmark that systematically replicates and evaluates the full social survey pipeline using LLMs. It defines four tasks aligned with key survey stages: social role modeling, semi-structured interview modeling, attitude stance modeling and survey response modeling. It also provides task-specific evaluation metrics to assess alignment fidelity, consistency, and fairness at both individual and group levels, with a focus on demographic diversity. To support AlignSurvey, we construct a multi-tiered dataset architecture: (i) the Social Foundation Corpus, a cross-national resource with 44K+ interview dialogues and 400K+ structured survey records; and (ii) a suite of Entire-Pipeline Survey Datasets, including the expert-annotated AlignSurvey-Expert (ASE) and two nationally representative surveys for cross-cultural evaluation. We release the SurveyLM family, obtained through two-stage fine-tuning of open-source LLMs, and offer reference models for evaluating domain-specific alignment. All datasets, models, and tools are available at github and huggingface to support transparent and socially responsible research.

Qingqing Liang, Chunyang Wang, Peiwei Xia, Yanan Zeng, Xin Liu, Xuesong Lu, Aoying Zhou

Educational recommendation systems have been a fundamental component for alleviating learning disorientation in self-paced learning. While existing studies mainly leverage cognitive theories to guide learning motivation modeling, they critically overlook the role of social influences. Through empirical analysis, we identify social homophily as an additional driver of learning behaviors, i.e., learners tend to adopt resources validated by their social cohort. However, two challenges impede effective social homophily modeling: (1) the absence and sparsity of predefined social relations in online education, and (2) the deep entanglement of social homophily with cognitive homophily in behavioral data. To tackle these challenges, we propose a graph-based framework EdGCL that explicitly disentangles social homophily and cognitive homophily. EdGCL infers implicit social relations from learners' social behaviors and encodes them via a graph transformer, generating social-view representations. Simultaneously, it constructs a heterogeneous learning graph to model cognitive homophily, which is enhanced by a type-aware aggregator and cognitive diagnosis loss. To ensure the semantic distinctiveness of dual-view homophily modeling, a cross-view contrastive disentanglement mechanism is designed to pull intra-view representations closer while pushing inter-view representations away. Evaluation on two real-world educational datasets demonstrates the superior recommendation performance of EdGCL, highlighting the necessity of dual homophily modeling for understanding the motivations behind learning behaviors.

Shuaibo Li, Laixin Zhang, Wei Ma, Jianwei Guo, Shibiao Xu, Zhijie Qiu, Hongbin Zha

Detecting AI-generated images remains a persistent challenge, as existing detectors often struggle to generalize to forgeries produced by previously unseen generative models. This generalization gap mainly stems from entanglement with semantic content and overfitting to model-specific artifacts. Moreover, many state-of-the-art methods rely on large pre-trained backbones or computationally intensive pipelines, which limit their applicability in real-world, resource-constrained environments. We propose RealNet, a lightweight and unsupervised framework that constructs a disentangled, forgery-aware representation space using only real images. RealNet first extracts semantic-agnostic representations through a dual adversarial denoising mechanism, producing compact features with low intra-class variance. These representations are then perturbed in feature space to generate pseudo-negative samples, which are combined with the original real features to train a lightweight discriminator, enabling robust detection without any dependence on synthetic images during training. Comprehensive evaluations across GAN, diffusion, and emerging VAR-based paradigms demonstrate that RealNet achieves superior cross-model generalization and robustness. RealNet surpasses previous state-of-the-art approaches by 4.51% in accuracy and 3.93% in average precision, while maintaining significantly lower computational cost. Furthermore, we introduce a medically relevant synthetic image dataset and show RealNet remains effective under severe distribution shifts, highlighting its potential for deployment in high-stakes real-world scenarios. Together, these advantages position RealNet as a practical, scalable and socially impactful solution for robust AI-generated image detection.

Ruohan Li, Zhihao Wang, Xiaowei Jia, Gengchen Mai, Lei Ma, George C. Hurtt, Quan Shen, Zhili Li, Yiqun Xie

Terrestrial ecosystems constitute a major component of the global carbon sink and play a critical role in regulating the global carbon cycle. Although process-based models such as the Ecosystem Demography (ED) model are widely used to simulate these dynamics and widely adopted in research and applications, they remain computationally intensive and are not well suited for large-scale (e.g., global) projections at high spatial and temporal resolution, or under wide-range of future scenarios. AI-based emulators of process-based physical models have emerged as promising ways to accelerate the computation. However, there are several challenges in developing emulators for ecosystem processes, including error accumulation over long sequences, single-step initial conditions, and high-dimensional environmental conditions. Existing works often rely on time-series patterns in look-back windows, which are not well-suited for the problem with single-step initial conditions. Moreover, they often do not consider uncertainty, making it hard to know when the approximations are highly confident and when the results may need to be updated, e.g., by the process-based models. To address these limitations, we introduce EcoDiffusion, a conditional diffusion framework tailored for ecosystem dynamics emulation. We evaluated EcoDiffusion at locations distributed worldwide under different scenarios and showed that it demonstrated significant improvements over existing models.

Naiyi Li, Zihui Ma, Runlong Yu, Lingyao Li

Digital Twins (DTs) offer powerful tools for managing complex infrastructure systems, but their effectiveness is often limited by challenges in integrating unstructured knowledge. Recent advances in Large Language Models (LLMs) bring new potential to address this gap, with strong abilities in extracting and organizing diverse textual information. We therefore propose LSDTs (LLM-Augmented Semantic Digital Twins), a framework that helps LLMs extract planning knowledge from unstructured documents like environmental regulations and technical guidelines, and organize it into a formal ontology. This ontology forms a semantic layer that powers a digital twin—a virtual model of the physical system—allowing it to simulate realistic, regulation-aware planning scenarios. We evaluate LSDTs through a case study of offshore wind farm planning in Maryland, including its application during Hurricane Sandy. Results demonstrate that LSDTs support interpretable, regulation-aware layout optimization, enable high-fidelity simulation, and enhance adaptability in infrastructure planning. This work shows the potential of combining generative AI with digital twins to support complex, knowledge-driven planning tasks.

Hao Li, Haotian Chen, Ruoyuan Gong, Juanjuan Wang, Hao Jiang

Redistricting plays a central role in shaping how votes are translated into political power. While existing computational methods primarily aim to generate large ensembles of legally valid districting plans, they often neglect the strategic dynamics involved in the selection process. This oversight creates opportunities for partisan actors to cherry-pick maps that, while technically compliant, are politically advantageous. Simply satisfying formal constraints does not ensure fairness when the selection process itself can be manipulated. We propose Agentmandering, a framework that reimagines redistricting as a turn-based negotiation between two agents representing opposing political interests. Drawing inspiration from game-theoretic ideas, particularly the Choose-and-Freeze protocol, our method embeds strategic interaction into the redistricting process via large language model (LLM) agents. Agents alternate between selecting and freezing districts from a small set of candidate maps, gradually partitioning the state through constrained and interpretable choices. Evaluation on post-2020 U.S. Census data across all states shows that Agentmandering significantly reduces partisan bias and unfairness, while achieving 2 to 3 orders of magnitude lower variance than standard baselines. These results demonstrate both fairness and stability, especially in swing-state scenarios.

Bingxuan Li, Pengyi Shi, Amy R Ward

Predictive modeling in high-stakes domains often suffers from limited observed features due to ethical and practical constraints. To address this challenge, we propose a novel approach that formulates latent feature mining as a text-to-text propositional logic reasoning task, facilitating domain knowledge integration and improving the interpretability of latent features. We design FLAME, a domain knowledge-augmented reasoning framework for latent feature mining, offering an efficient training paradigm to strengthen the domain-specific reasoning capabilities of large language models (LLMs) for latent feature extraction. The goal of our framework is to augment observed features with inferred latent features, enhancing the performance of predictive models in downstream machine learning tasks. We validate our approach through two case studies: (1) the criminal justice system, where data collection is ethically challenging and inherently limited, and (2) the healthcare domain, where patient privacy concerns and the complexity of medical data restrict comprehensive feature collection. Experimental results demonstrate that the inferred latent features significantly enhance the performance of downstream classifiers by over 10%.

Matan Levy, Itay Segev, Alexander Tuisov, Sarah Keren

The objective of this study is to advance the optimization of hybrid electricity markets using multi-agent reinforcement learning (MARL). The transition from centralized systems to public–private models introduces significant challenges, including the emergence of independent market players and the increasing integration of renewable energy sources (RESs). These challenges are further intensified by rapidly shifting demand patterns, driven both by energy-intensive data centers and AI inference workloads, as well as by political and societal instabilities. To address these complexities, we develop a formal model of market participants’ behavior and propose a MARL-based framework for optimizing system operator strategies. This framework incorporates dynamic pricing and dispatch scheduling to minimize operational costs, maintain grid stability, and align market incentives. We also present a new, adaptable simulation environment compatible with state-of-the-art MARL methods. Empirical evaluations in increasingly complex scenarios demonstrate the effectiveness of our approach in capturing the dynamic and decentralized nature of modern electricity markets.

Yonggeon Lee, Jibin Hwang, Alfred Malengo Kondoro, Juhyun Song, Youngtae Noh

Electric vehicles (EVs) are key to sustainable mobility, yet their lithium-ion batteries (LIBs) degrade more rapidly under prolonged high states of charge (SOC). This can be mitigated by delaying full charging DFC until just before departure, which requires accurate prediction of user departure times. In this work, we propose Transformer-based real-time-to-event (TTE) model for accurate EV departure prediction. Our approach represents each day as a TTE sequence by discretizing time into grid-based tokens. Unlike previous methods primarily dependent on temporal dependency from historical patterns, our method leverages streaming contextual information to predict departures. Evaluation on a real-world study involving 93 users and passive smartphone data demonstrates that our method effectively captures irregular departure patterns within individual routines, outperforming baseline models. These results highlight the potential for practical deployment of the DFC algorithm and its contribution to sustainable transportation systems.

Sumin Lee, Sungwon Park, Jeasurk Yang, Jihee Kim, Meeyoung Cha

Satellite-based slum segmentation holds significant promise in generating global estimates of urban poverty. However, the morphological heterogeneity of informal settlements presents a major challenge, hindering the ability of models trained on specific regions to generalize effectively to unseen locations. To address this, we introduce a large-scale high-resolution dataset and propose GRAM (Generalized Region-Aware Mixture-of-Experts), a two-phase test-time adaptation framework that enables robust slum segmentation without requiring labeled data from target regions. We compile a million-scale satellite imagery dataset from 12 cities across four continents for source training. Using this dataset, the model employs a Mixture-of-Experts architecture to capture region-specific slum characteristics while learning universal features through a shared backbone. During adaptation, prediction consistency across experts filters out unreliable pseudo-labels, allowing the model to generalize effectively to previously unseen regions. GRAM outperforms state-of-the-art baselines in low-resource settings such as African cities, offering a scalable and label-efficient solution for global slum mapping and data-driven urban planning.

Sangyub Lee, Heedou Kim, Hyeoncheol Kim

The use of Large Language Models (LLMs) in police opera- tions is growing, yet an evaluation framework tailored to po- lice operations remains absent. While LLM’s responses may not always be legally “incorrect”, their unverified use still can lead to severe issues such as unlawful arrests and improper evidence collection. To address this, we propose PAS (Po- lice Action Scenarios), a systematic framework covering the entire evaluation process. Applying this framework, we con- structed a novel QA dataset from over 8,000 official docu- ments and established key metrics validated through statis- tical analysis with police expert judgements. Experimental results show that commercial LLMs struggle with our new police-related tasks, particularly in providing fact-based rec- ommendations. This study highlights the necessity of an ex- pandable evaluation framework to ensure reliable AI-driven police operations. We release our data and prompt template.

Chung Peng Lee, Rachel Hong, Harry H. Jiang, Aster Plotnik, William Agnew, Jamie Heather Morgenstern

The internet has become the main source of data to train modern text-to-image or vision-language models, yet it is increasingly unclear whether web-scale data collection practices for training AI systems adequately respect data owners' wishes. Ignoring the owner's indication of consent around data usage not only raises ethical concerns but also has recently been elevated into lawsuits around copyright infringement cases. In this work, we aim to reveal information about data owners' consent to AI scraping and training, and study how it's expressed in DataComp, a popular dataset of 12.8 billion text-image pairs. We examine both the sample-level information, including the copyright notice, watermarking, and metadata, and the web-domain-level information, such as a site's Terms of Service (ToS) and Robots Exclusion Protocol. We estimate at least 122M of samples exhibit some indication of copyright notice in CommonPool, and find that 60% of the samples in the top 50 domains come from websites with ToS that prohibit scraping. Furthermore, we estimate 9-13% with 95% confidence interval of samples from CommonPool to contain watermarks, where existing watermark detection methods fail to capture them in high fidelity. Our holistic methods and findings show that data owners rely on various channels to convey data consent, of which current AI data collection pipelines do not entirely respect. These findings highlight the limitations of the current dataset curation/release practice and the need for a unified data consent framework taking AI purposes into consideration.

Hou Hei Lam, Jiangjie Qiu, Xiuyuan Hu, Wentao Li, Fankun Zeng, Siwei Fu, Hao Zhang, Xiaonan Wang

Organic photovoltaic (OPV) materials offer a promising pathway for sustainable energy generation. However, their development is hindered by the challenge of identifying high-performance donor-acceptor pairs with optimal power conversion efficiencies (PCEs). Most existing design strategies focus exclusively on either the donor or the acceptor, rather than employing a unified model capable of designing both components. In this work, we introduce a dual-pronged machine learning framework for OPV discovery, integrating predictive modeling and generative molecular design. In this study, we propose the newly curated Organic Photovoltaic Donor-Acceptor Dataset (OPV²D), the largest of its kind, comprising 2,000 experimentally characterized donor-acceptor pairs. This dataset serves as a comprehensive foundation for model training and evaluation. To enable accurate property prediction in organic photovoltaic (OPV) materials, we first introduce the Organic Photovoltaic Classifier (OPVC) to predict the likelihood that a given material exhibits OPV behavior. Complementing this, we develop a hierarchical graph neural network framework that integrates multi-task learning and cross-modal donor–acceptor interaction modeling. This framework includes the Molecular Orbital Energy Estimator (MOE²) for predicting the highest occupied molecular orbital–lowest unoccupied molecular orbital (HOMO–LUMO) energy levels, and the Photovoltaic Performance Predictor (P³) for estimating power conversion efficiency (PCE). In addition, we introduce the Material Generative Pretrained Transformer (MatGPT) to generate synthetically accessible organic semiconductors. Building on this, we propose a reinforcement learning strategy with three-objective policy optimization that guides molecular generation while preserving chemical validity. By bridging molecular representation learning with device performance prediction, our framework advances computational OPV material discovery.

Deuksin Kwon, Kaleen Shrestha, Bin Han, Spencer Lin, James Hale, Jonathan Gratch, Maja Mataric, Gale M. Lucas

Large language models (LLMs) are increasingly used to simulate human behavior in social settings such as legal mediation, negotiation, and dispute resolution. However, it remains unclear whether these simulations reproduce the personality–behavior patterns observed in humans. Human personality, for instance, shapes how individuals navigate social interactions, including strategic choices and behaviors in emotionally charged interactions. This raises the question: Can LLMs, when prompted with personality traits, reproduce personality-driven differences in human conflict behavior? To explore this, we introduce an evaluation framework that enables direct comparison of human-human and LLM-LLM behaviors in dispute resolution dialogues with respect to Big Five Inventory (BFI) personality traits. This framework provides a set of interpretable metrics related to strategic behavior and conflict outcomes. We additionally contribute a novel dataset creation methodology for LLM dispute resolution dialogues with matched scenarios and personality traits with respect to human conversations. Finally, we demonstrate the use of our evaluation framework with three contemporary closed-source LLMs and show significant divergences in how personality manifests in conflict across different LLMs compared to human data, challenging the assumption that personality-prompted agents can serve as reliable behavioral proxies in socially impactful applications. Our work highlights the need for psychological grounding and validation in AI simulations before real-world use.

Chenchen Kuai, Chenhao Wu, Yang Zhou, Bruce Wang, Tianbao Yang, Zhengzhong Tu, Zihao Li, Yunlong Zhang

As tropical cyclones intensify and track forecasts become increasingly uncertain, U.S. ports face heightened supply-chain risk under extreme weather conditions. Port operators need to rapidly synthesize diverse multimodal forecast products, such as probabilistic wind maps, track cones, and official advisories, into clear, actionable guidance as cyclones approach. Multimodal large language models (MLLMs) offer a powerful means to integrate these heterogeneous data sources alongside broader contextual knowledge, yet their accuracy and reliability in the specific context of port cyclone preparedness have not been rigorously evaluated. To fill this gap, we introduce CyPortQA, the first multimodal benchmark tailored to port operations under cyclone threat. CyPortQA assembles 2,917 real-world disruption scenarios from 2015 through 2023, spanning 145 U.S. principal ports and 90 named storms. Each scenario fuses multi-source data (i.e., tropical cyclone products, port operational impact records, and port condition bulletins) and is expanded through an automated pipeline into 117,178 structured question–answer pairs. Using this benchmark, we conduct extensive experiments on diverse MLLMs, including both open-source and proprietary model. MLLMs demonstrate great potential in situation understanding but still face considerable challenges in reasoning tasks, including potential impact estimation and decision reasoning.