Large Language Models (LLMs) enable Web-scale multilingual content analysis but face critical challenges in scaling to long-tail languages and ensuring robustness. Current research is split between two isolated trajectories: a Macro-Paradigm (system-level engineering) and a Micro-Paradigm (internal model intervention). We argue that a true Web-scale solution requires their systematic fusion, balancing large-scale data processing with fine-grained model control. We introduce the Control-Tower Framework (CTF), a novel methodology designed to systematically enhance powerful, pre-trained base models. Inspired by control-theoretic ideas, CTF transforms a base model into a controllable analysis engine via three synergistic stages: (1) Micro-enhanced pre-training that injects linguistic priors (e.g., syntax) to build a robust semantic foundation; (2) a control-inspired fine-tuning stage where a heuristic dynamic feedback loop, driven by micro-level error signals (e.g., knowledge editing loss), actively adjusts the macro-scale learning curriculum; and (3) Macro-optimized inference using Minimum Bayes Risk (MBR) decoding to enhance robustness on noisy user-generated content (UGC). Extensive experiments show that CTF surpasses the leading open-weights model, Tower+ 9B FT, by a substantial margin of +2.18 XCOMET-XXL on low-resource languages (WMT24++). Crucially, CTF unlocks large-scale cross-lingual Web mining by converting unstructured Web text into machine-analyzable assets. We evidence this with substantial gains across both document-level (on MARC) and aspect-based (on SemEval-2016) sentiment analysis tasks. Our work offers a practical pathway toward building more reliable, scalable, and controllable global information ecosystems.
论文检索
输入标题、作者或关键词，从 100,903 篇学术成果中精准定位
Online Time Series Forecasting (OTSF) requires models to continuously adapt to concept drift. However, existing methods often treat concept drift as a monolithic phenomenon. To address this limitation, we first redefine concept drift by categorizing it into two distinct types: Recurring Drift, where previously seen patterns reappear, and Emergent Drift, where entirely new patterns emerge. We then propose DynaME (Dynamic Multi-period Experts), a novel hybrid framework designed to effectively address this dual nature of drift. For Recurring Drift, DynaME employs a committee of specialized experts that are dynamically fitted to the most relevant historical periodic patterns at each time step. For Emergent Drift, the framework detects high-uncertainty scenarios and shifts reliance to a stable, general expert. Extensive experiments on several benchmark datasets and backbones demonstrate that DynaME effectively adapts to both concept drifts and significantly outperforms existing baselines. Our source code is available at https://github.com/shhong97/DynaME.
Understanding how media rhetoric shapes audience engagement is crucial in the attention economy. This study examines how moral-emotional framing by mainstream news channels on YouTube influences user behavior across Korea and the United States. To capture the platform's multimodal nature, combining thumbnail images and video titles, we develop a multimodal moral emotion classifier by fine-tuning a vision–language model. The model is trained on human-annotated multimodal datasets in both languages and applied to approximately 400,000 videos from major news outlets. We analyze three engagement levels (views, likes, and comments), representing increasing degrees of commitment. The results show that other-condemning rhetoric—expressions of moral outrage that criticize others' morality—consistently increases all forms of engagement across cultures, with effect sizes strengthening from passive viewing to active commenting. These findings suggest that moral outrage is a particularly effective emotional strategy, attracting not only attention but also active participation. We discuss concerns about the potential misuse of other-condemning rhetoric, as such practices may deepen polarization by reinforcing in-group/out-group divisions. To facilitate future research and ensure reproducibility, we publicly release our Korean and English multimodal moral emotion classifiers.
With the rapid advancement of large language models (LLMs), a diverse ecosystem of models with different scales and domain specializations has emerged, including LLM-based web agents and online multi-LLM server systems. LLM routing, which opportunistically leverages this diversity to balance response quality and computational cost, has become a central problem in optimizing the performance of LLM serving systems. We propose the Budget-Adaptive Router (BARouter), a budget-adaptive online routing framework for LLM serving systems. BARouter dynamically adjusts its routing policy for incoming queries based on the estimated response quality, query costs, and real-time budget consumption, enabling seamless adaptation to varying initial budgets and evolving input distributions without manual hyperparameter tuning. Theoretically, we prove that BARouter achieves sublinear regret over the time horizon T. The extensive experiments show that BARouter effectively allocates the right budget to the queries at the right time, consistently outperforming baseline algorithms, and remains robust to varying budget levels and shifting query distributions.
With the rapid advancement of large language models (LLMs), producing realistic fake news has become increasingly effortless, challenging existing detection methods that rely on lexical and syntactic patterns. To address this, we shift our focus to the generation process and analyze how malicious prompts manipulate model outputs. We construct pairs of LLM-generated real and fake news and apply malicious prompts to reconstruct them as fake. By comparing the original-token generation probabilities recorded during reconstruction, we observe a consistent statistical divergence: tokens from real news tend to have lower reconstruction likelihoods than those from fake news. We define this distributional divergence as linguistic fingerprint. Building on this insight, we propose LIFE (Linguistic Fingerprints Extraction), a novel detection framework that reconstructs token-level probability distributions guided by malicious prompts to capture these discriminative linguistic patterns. To fully exploit the extracted fingerprints, LIFE further introduces a key-fragment amplification module that adaptively identifies and accentuates the most distinctive linguistic fragments, thereby enhancing detection reliability across diverse prompting scenarios. Extensive experiments demonstrate that LIFE achieves state-of-the-art performance in detecting LLM-generated fake news while maintaining strong generalization to human-LLM mixed cases. The code is available.
The rise of micro-videos has reshaped how misinformation spreads, amplifying its speed, reach, and impact on public trust. Existing benchmarks typically focus on a single deception type, overlooking the diversity of real-world cases that involve multimodal manipulation, AI-generated content, cognitive bias, and out-of-context reuse. Meanwhile, most detection models lack fine-grained attribution, limiting interpretability and practical utility. To address these gaps, we introduce WildFakeBench, a large-scale benchmark of over 10,000 real-world micro-videos covering diverse misinformation types and sources, each annotated with expert-defined attribution labels. Building on this foundation, we develop FakeAgent, a Delphi-inspired multi-agent reasoning framework that integrates multimodal understanding with external evidence for attribution-grounded analysis. FakeAgent jointly analyzes content and retrieved evidence to identify manipulation, recognize cognitive and AI-generated patterns, and detect out-of-context misinformation. Extensive experiments show that FakeAgent consistently outperforms existing MLLMs across all misinformation types, while WildFakeBench provides a realistic and challenging testbed for advancing explainable micro-video misinformation detection.. Data and code are available at: https://github.com/Aiyistan/FakeAgent.
Time series forecasting underpins many real-world services. Recent trends have focused on foundation models inspired by the paradigm of large language models, which rely on large volumes of centralized time-series data across diverse domains. However, such approaches raise significant concerns regarding data privacy. Federated learning (FL) has emerged as a promising paradigm for training unified time-series models using isolated datasets distributed across multiple clients. Nevertheless, existing FL methods face two critical challenges: heterogeneous variables and heterogeneous temporal correlations. To address these issues, we propose FedRMamba, a personalized federated forecasting framework built entirely from Mamba state-space blocks. Each client adopts a residual-coupled architecture, where a global frequency-aware Mamba module captures the common low-frequency structures shared across different variables, while a local patch-wise Mamba module learns personalized high-frequency patterns within the multivariate context. To clearly separate these responsibilities, we introduce a frequency-aware supervision that aligns the global path with low-frequency components and the local path with high-frequency residuals. Additionally, we design a gated fusion mechanism that dynamically combines the low-frequency and high-frequency components for improved prediction. We conduct extensive experiments to evaluate the performance of our proposed framework, demonstrating its effectiveness in handling heterogeneous data in federated settings.
As a key to accessing research impact, citation dynamics underpins research evaluation, scholarly recommendation, and the study of knowledge diffusion. Citation prediction is particularly critical for newborn papers, where early assessment must be performed without citation signals and under highly long-tailed distributions. We identify two key research gaps: (i) insufficient modeling of implicit factors of scientific impact, leading to reliance on coarse proxies; and (ii) a lack of bias-aware learning that can deliver stable predictions on lowly cited papers. We address these gaps by proposing a Bias-Aware Citation Prediction Framework, which combines multi-agent feature extraction with robust graph representation learning. First, a multi-agent × graph co-learning module derives fine-grained, interpretable signals, such as reproducibility, collaboration network, and text quality, from metadata and external resources, and fuses them with heterogeneous-network embeddings to provide rich supervision even in the absence of early citation signals. Second, we incorporate a set of robust mechanisms: a two-stage forward process that routes explicit factors through an intermediate exposure estimate, GroupDRO to optimize worst-case group risk across environments, and a regularization head that performs what-if analyses on controllable factors under monotonicity and smoothness constraints. Comprehensive experiments on two real-world datasets demonstrate the effectiveness of our proposed model. Specifically, our model achieves around a 13% reduction in error metrics (MALE and RMSLE) and a notable 5.5% improvement in the ranking metric (NDCG) over the baseline methods.
Urban region profiling, the task of characterizing geographical areas, is crucial for urban planning and resource allocation. However, existing research in this domain faces two significant limitations. First, most methods are confined to single-task prediction, failing to capture the interconnected, multi-faceted nature of urban environments where numerous indicators are deeply correlated. Second, the field lacks a standardized experimental benchmark, which severely impedes fair comparison and reproducible progress. To address these challenges, we first establish a comprehensive benchmark for multi-task urban region profiling, featuring multi-modal features and a diverse set of strong baselines to ensure a fair and rigorous evaluation environment. Concurrently, we propose UrbanMoE, the first sparse multi-modal, multi-expert framework specifically architected to solve the multi-task challenge. Leveraging a sparse Mixture-of-Experts architecture, it dynamically routes multi-modal features to specialized sub-networks, enabling the simultaneous prediction of diverse urban indicators. We conduct extensive experiments on three real-world datasets within our benchmark, where UrbanMoE consistently demonstrates superior performance over all baselines. Further in-depth analysis validates the efficacy and efficiency of our approach, setting a new state-of-the-art and providing the community with a valuable tool for future research in urban analytics.
Retrieval-Augmented Generation (RAG)-enhanced LLM systems, while powerful, introduce substantial inference costs due to the inclusion of an extra multi-stage pipeline that dynamically retrieves and synthesizes information from external knowledge sources. This high operational cost exposes a critical vulnerability to Inference Cost Attacks (ICAs). However, existing ICAs often rely on the impractical assumption of direct prompt manipulation. We argue that a more feasible and potent threat to RAG-enhanced LLM systems arises from poisoning external knowledge bases (e.g., web knowledge from the Internet). In this work, we introduce the Retrieval-Augmented Inference Cost Attack (RA-ICA), a novel attacking paradigm that targets the computational cost of RAG-enhanced LLM systems by injecting malicious documents into external knowledge corpus. To operationalize this attack, we propose Computational Resource Exhaustion via External Poisoning (CREEP), a novel framework that leverages LLM agents to automatically craft malicious documents that are both semantically relevant for retrieval and potent for inducing an abnormal increase in token consumption during the inference phase. To enhance the attack's effectiveness, we introduce Memory-Augmented Group Relative Policy Optimization (MA-GRPO), a novel reinforcement learning algorithm that fine-tunes the agents by learning from a dynamic memory of historical best adversarial documents. Extensive experiments across three real-world datasets demonstrate that RA-ICA increases token consumption by up to 13.12 times with an over 90% success rate, without degrading the integrity of the generated answer.
Real-world multimodal knowledge graphs (MKGs) are inherently heterogeneous, modeling entities that are associated with diverse modalities. Traditional knowledge graph embedding (KGE) methods excel at learning continuous representations of entities and relations, yet they are typically designed for unimodal settings. Recent approaches extend KGE to multimodal settings but remain constrained, often processing modalities in isolation, resulting in weak cross-modal alignment, and relying on simplistic assumptions such as uniform modality availability across entities. Vision--Language Models (VLMs) offer a powerful way to align diverse modalities within a shared embedding space. We propose Vision--Language Knowledge Graph Embeddings (VL-KGE), a framework that integrates cross-modal alignment from VLMs with structured relational modeling to learn unified multimodal representations of knowledge graphs. Experiments on WN9-IMG and two novel fine art MKGs, WikiArt-MKG-v1 and WikiArt-MKG-v2, demonstrate that VL-KGE consistently improves over traditional unimodal and multimodal KGE methods in link prediction tasks. Our results highlight the value of VLMs for multimodal KGE, enabling more robust and structured reasoning over large-scale heterogeneous knowledge graphs.
Federated learning offers a promising paradigm for privacy-preserving traffic prediction, yet its performance is often challenged by the non-identically and independently distributed (non-IID) nature of decentralized traffic data. Existing federated methods frequently struggle with this data heterogeneity, typically entangling globally shared patterns with client-specific local dynamics within a single representation. In this work, we postulate that this heterogeneity stems from the entanglement of two distinct generative sources: client-specific localized dynamics and cross-client global spatial-temporal patterns. Motivated by this perspective, we introduce FedDis, a novel framework that, to the best of our knowledge, is the first to leverage causal disentanglement for federated spatial-temporal prediction. Architecturally, FedDis comprises a dual-branch design wherein a Personalized Bank learns to capture client-specific factors, while a Global Pattern Bank distills common knowledge. This separation enables robust cross-client knowledge transfer while preserving high adaptability to unique local environments. Crucially, a mutual information minimization objective is employed to enforce informational orthogonality between the two branches, thereby ensuring effective disentanglement. Comprehensive experiments conducted on four real-world benchmark datasets demonstrate that FedDis consistently achieves state-of-the-art performance, promising efficiency, and superior expandability. To ease the reproducibility, we release our implementation code online. https://github.com/Jlu-zcy/www2026_FedDis.
Large Language Models (LLMs) are increasingly relying on web crawling to stay up to date and accurately answer user queries. These crawlers are expected to honor robots.txt files, which govern automated access. In this study, for the first time, we investigate whether reputable news websites and misinformation sites differ in how they configure these files, particularly in relation to AI crawlers. Analyzing a curated dataset, we find a stark contrast: 60.0% of reputable sites disallow at least one AI crawler, compared to just 9.1% of misinformation sites in their robots.txt files. Reputable sites forbid an average of 15.5 AI user agents, while misinformation sites prohibit fewer than one. We then measure active blocking behavior, where websites refuse to return content when HTTP requests include AI crawler user agents, and disclose that both categories of websites utilize it. Notably, the behavior of reputable news websites in this regard aligns more closely with their declared robots.txt directive than with that of misinformation websites. Finally, our longitudinal analysis reveals that this gap has widened over time, with AI-blocking by reputable websites via robots.txt increasing from 23% in September 2023 to nearly 60% by May 2025. Our findings suggest a growing asymmetry in content accessibility, as reflected in robots.txt directives, that may shape the training data available to LLMs, raising essential questions for web transparency, data ethics, and the future of AI training practices.
Time series forecasting is crucial for the World Wide Web and represents a core technical challenge in ensuring the stable and efficient operation of modern web services, such as intelligent transportation and website throughput. However, we have found that existing methods typically employ a strategy of embedding each time step as an independent token. This paradigm introduces a fundamental information bottleneck when processing long sequences, the root cause of which is that independent token embedding destroys a crucial structure within the sequence—what we term as multi-offset temporal correlation. This refers to the fine-grained dependencies embedded within the sequence that span across different time steps, which is especially prevalent in regular Web data. To fundamentally address this issue, we propose a new perspective on time series embedding. We provide an upper bound on the approximate reconstruction performance of token embedding, which guides our design of a concise yet effective Multi-Offset Time Embedding (MOTE) method to mitigate the performance degradation caused by standard token embedding. Furthermore, our MOTE can be integrated into various existing models and serve as a universal building block. Based on this paradigm, we further design a novel forecasting architecture named Time-TK. This architecture first utilizes a Multi-Offset Interactive KAN (MI-KAN) to learn and represent specific temporal patterns among multiple offset sub-sequences. Subsequently, it employs an efficient Multi-Offset Temporal Interaction mechanism (MOTI) to effectively capture the complex dependencies between these sub-sequences, achieving global information integration. Extensive experiments on 14 real-world benchmark datasets, covering domains such as traffic flow and BTC/USDT throughput, demonstrate that Time-TK significantly outperforms all baseline models, achieving state-of-the-art forecasting accuracy.
With the rapid development of web services, large amounts of time series data are generated and accumulated across various domains such as finance, healthcare, and online platforms. As such data often co-evolves with multiple variables interacting with each other, estimating the time-varying dependencies between variables (i.e., the dynamic network structure) has become crucial for accurate modeling. However, real-world data is often represented as tensor time series with multiple modes, resulting in large, entangled networks that are hard to interpret and computationally intensive to estimate. In this paper, we propose Kronecker Time-Varying Graphical Lasso (KTVGL), a method designed for modeling tensor time series. Our approach estimates mode-specific dynamic networks in a Kronecker product form, thereby avoiding overly complex entangled structures and producing interpretable modeling results. Moreover, the partitioned network structure prevents the exponential growth of computational time with data dimension. In addition, our method can be extended to stream algorithms, making the computational time independent of the sequence length. Experiments on synthetic data show that the proposed method achieves higher edge estimation accuracy than existing methods while requiring less computation time. To further demonstrate its practical value, we also present a case study using real-world data. Our source code and datasets are available at https://github.com/Higashiguchi-Shingo/KTVGL.
Modern web applications—ranging from personalized recommendation to real-time fraud detection—rely on AI models to deliver timely and personalized services, yet the underlying user interaction data arrives as massive and evolving streams. Stream Learning (SL) offers a natural paradigm for building adaptive models, but it struggles with challenges such as redundant training data and catastrophic forgetting, which can undermine long-term predictive performance. To address these issues, recent studies have explored data selection strategies like coreset selection and buffer update, typically implemented through rule-based or model-based methods. However, fixed selection rules hinder the adaptability of rule-based approaches to changing data distributions, while model-based methods often depend on costly per-sample gradients, leading to throttled updates and reduced coverage of informative samples. In this paper, we propose StreamFP, a lightweight SL framework that introduces fingerprints-a set of compact, learnable parameter vectors that summarize the model state. Fingerprints compute similarity scores to jointly guide coreset selection and buffer update, prioritizing informative incoming samples while retaining representative historical ones. A lightweight fingerprint attunement plugin further calibrates fingerprints using pre-trained ViT attention with negligible overhead, thereby improving accuracy while mitigating forgetting. Extensive experiments demonstrate that StreamFP consistently achieves superior accuracy and efficiency compared with state-of-the-art methods across diverse real-world datasets and varying data arrival rates.
Modeling asynchronous event sequences is crucial in numerous real-world applications such as healthcare monitoring, financial transaction analysis, and so on. Traditional temporal point processes, including Hawkes Processes, often fail to capture complex dependencies due to their parametric limitations. While neural approaches like RNNs and Transformers have improved flexibility, they struggle with computational inefficiency, and attention saturation. In this paper, we introduce the Mamba Hawkes Process (MHP), the first framework to integrate selective state space model (Mamba) with temporal point processes. MHP leverages time-varying state transitions and input-dependent gating to efficiently encode event history and capture long-term dependencies with linear complexity. Importantly, we provide theoretical guarantees showing that MHP generalizes both classical multi-exponential Hawkes processes and exponential-decay gated RNNs, underscoring its expressive power and theoretical soundness. To address the inherent constraints of pure state space models in handling heterogeneous event interactions, we further develop Adaptive Mamba Hawkes Process (A-MHP) that incorporates two novel mechanisms: a Time-Scaling Mechanism that adaptively weights time intervals based on event type and history, and a Dual-Channel State Transition that adaptively processes event content and temporal dynamics for more refined state updates. Extensive experiments on synthetic and real-world datasets demonstrate that MHP and A-MHP consistently outperform state-of-the-art baselines in event prediction tasks, particularly in long-sequence scenarios. Our work establishes a scalable and theoretically grounded paradigm for event sequence modeling, with practical implications for predictive maintenance, anomaly detection, and dynamic system analysis. The code is available at https://github.com/Ethan-Shen-Individual-Lab/Mamba-Hawkes-Process.