We present Autoregressive Representation Alignment (ARRA), a new training framework that unlocks global-coherent text-to-image generation in autoregressive LLMs without architectural modifications. Different from prior works that require complex architectural redesigns, ARRA aligns LLM's hidden states with visual representations from external visual foundational models via a global visual alignment loss and a hybrid token, . This token enforces dual constraints: local next-token prediction and global semantic distillation, enabling LLMs to implicitly learn spatial and contextual coherence while retaining their original autoregressive paradigm. Extensive experiments validate ARRA's plug-and-play versatility. When training T2I LLMs from scratch, ARRA reduces FID by 16.6% (ImageNet), 12.0% (LAION-COCO) for autoregressive LLMs like LlamaGen, without modifying original architecture and inference mechanism. For training from text-generation-only LLMs, ARRA reduces FID by 25.5% (MIMIC-CXR), 8.8% (DeepEyeNet) for advanced LLMs like Chameleon. For domain adaptation, ARRA aligns general-purpose LLMs with specialized models (e.g., BioMedCLIP), achieving an 18.6% FID reduction over direct fine-tuning on medical imaging (MIMIC-CXR). These results demonstrate that training objective redesign, rather than architectural modifications, can resolve cross-modal global coherence challenges. ARRA offers a complementary paradigm for advancing autoregressive models.

Wentao Ge, Shunian Chen, Hardy Chen, Nuo Chen, Junying Chen, Zhihong Chen, Wenya Xie, Shuo Yan, Chenghao Zhu, Ziyue Lin, Dingjie Song, Xidong Wang, Anningzhe Gao, Zhang Zhiyi, Jianquan Li, Xiang Wan, Benyou Wang. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

The increasing popularity of Virtual Reality (VR) has provided a new medium for narrating spatiotemporal data stories. Compared to traditional 2D environments, telling spatiotemporal stories in VR holds the promise of offering a more immersive and engaging experience for audiences. However, when creating VR spatiotemporal data stories, story creators may feel overwhelmed by the numerous narrative elements involved and there is a disconnect between the creation and viewing environments. To address these challenges, we designed a hierarchical narrative structure with three levels: Story Line, Story Piece, and Viewpoint. We then introduce ST2VR, an interactive authoring system that supports the creation of spatiotemporal data stories within an immersive environment. The system features two types of interactive interfaces to support the organization of the overall narrative as well as the immersive design of story details. A use case demonstrates the process of authoring VR spatiotemporal data stories using ST2VR, and a user study evaluates the system’s efficiency and effectiveness.

Understanding user behaviors on social media has garnered significant scholarly attention, enhancing our comprehension of how virtual platforms impact society and empowering decision-makers. Simulating social media behaviors provides a robust tool for capturing the patterns of social media behaviors, testing hypotheses, and predicting the effects of various interventions, ultimately contributing to a deeper understanding of social media environments. Moreover, it can overcome difficulties associated with utilizing real data for analysis, such as data accessibility issues, ethical concerns, and the complexity of processing large and heterogeneous datasets. However, researchers and stakeholders need more flexible platforms to investigate different user behaviors by simulating different scenarios and characters, which is not possible yet. Therefore, this paper introduces SimSpark, an interactive system including simulation algorithms and interactive visual interfaces which is capable of creating small simulated social media platforms with customizable characters and social environments. We address three key challenges: generating believable behaviors, validating simulation results, and supporting interactive control for generation and results analysis. A simulation workflow is introduced to generate believable behaviors of agents by utilizing large language models. A visual interface enables real-time parameter adjustment and process monitoring for customizing generation settings. A set of visualizations and interactions are also designed to display the models' outputs for further analysis. Effectiveness is evaluated through case studies, quantitative simulation model assessments, and expert interviews.

Large Language Models (LLMs) have shown great potential in intelligent visualization systems, especially for domain-specific applications. Integrating LLMs into visualization systems presents challenges, and we categorize these challenges into three alignments: domain problems with LLMs, visualization with LLMs, and interaction with LLMs. To achieve these alignments, we propose a framework and outline a workflow to guide the application of fine-tuned LLMs to enhance visual interactions for domain-specific tasks. These alignment challenges are critical in education because of the need for an intelligent visualization system to support beginners' self-regulated learning. Therefore, we apply the framework to education and introduce Tailor-Mind, an interactive visualization system designed to facilitate self-regulated learning for artificial intelligence beginners. Drawing on insights from a preliminary study, we identify self-regulated learning tasks and fine-tuning objectives to guide visualization design and tuning data construction. Our focus on aligning visualization with fine-tuned LLM makes Tailor-Mind more like a personalized tutor. Tailor-Mind also supports interactive recommendations to help beginners better achieve their learning goals. Model performance evaluations and user studies confirm that Tailor-Mind improves the self-regulated learning experience, effectively validating the proposed framework.

With the increasing role of social media in information dissemination, effectively simulating and analyzing public event dynamics has become a key research focus. We present an interactive visual analysis system for simulating social media events using multi-agent models powered by large language models (LLMs). By modeling agents with diverse characteristics, the system explores how agents perceive information, adjust their emotions and stances, provide feedback, and influence the trajectory of events. The system integrates real-time interactive simulation with multi-perspective visualization, enabling users to investigate event trajectories and key influencing factors under varied configurations. Theoretical work standardizes agent attributes and interaction mechanisms, supporting realistic simulation of social media behaviors. Evaluation through indicators and case studies demonstrates the system’s effectiveness and adaptability, offering a novel tool for public event analysis across open social platforms.

For event analysis, the information from both before and after the event can be crucial in certain scenarios. By incorporating a contextualized perspective in event analysis, analysts can gain deeper insights from the events. We propose a contextualized visual analysis framework which enables the identification and interpretation of temporal patterns within and across multivariate events. The framework consists of a design of visual representation for multivariate event contexts, a data processing workflow to support the visualization, and a context-centered visual analysis system to facilitate the interactive exploration of temporal patterns. To demonstrate the applicability and effectiveness of our framework, we present case studies using real-world datasets from two different domains and an expert study conducted with experienced data analysts.

The exponential growth of the Internet of Things (IoT), driven by the increasing number of connected devices and sensors, is profoundly transforming many industries. This proliferation has been enabled by advances in storage and processing capabilities, facilitating the implementation of deep model. However, concerns surrounding privacy and substantial bandwidth required for data transmission arise from the centralized processing of data collected from numerous distributed devices. To address these issues, federated learning (FL) has emerged as a viable solution with a decentralized strategy. It ensures data privacy by allowing devices to locally train models while keeping their data localized, thus reducing centralization needs and concerns related to privacy and bandwidth. To mitigate the data heterogeneity problem in FL scenarios, we propose federated pattern extraction clustering (FedPEC) to cluster clients with similar data distributions. Compare to existing client clustering methods, FedPEC requires no additional data transmission and ensures great flexibility, scalability and privacy. Furthermore, we discuss the discrepancy between representation and data distribution in existing methods from the perspective of pattern representation. Based on this, we propose a variant of FedPEC as a solution, which has achieved excellent performance on multiple FL data sets.

This paper implements a vision-based moving target tracking system of quadrotors with visual-inertial localization in GNSS-denied indoor environments. We use the visual-inertial odometry to estimate the states of the UAV by minimizing visual and inertial residuals, and estimate the states of the target with extended Kalman Filter from visual detection. This research formulates the target tracking problem as optimization-based trajectory generation where a weighted sum cost function jointly penalizes the tracking error, the control cost of the trajectory and the trajectory length, while enforcing the safety and feasibility constraints. We present a strategy that represents the trajectory as piecewise Bézier curves using Bernstein polynomial basis. Due to the special properties of Bézier curves, the position of the entire trajectory and its derivatives can be directly bounded within the safe spaces, thus this facilitating the dynamics of the quadrotor. The proposed strategy can generate smooth and collision-free tracking trajectories and is time and space efficient. We conduct simulations and real-world experiments to validate the effectiveness of our system.