Table retrieval is the task of retrieving the most relevant tables from large-scale corpora given natural language queries. However, structural and semantic discrepancies between unstructured text and structured tables make embedding alignment particularly challenging. Recent methods such as QGpT attempt to enrich table semantics by generating synthetic queries, yet they still rely on coarse partial-table sampling and simple fusion strategies, which limit semantic diversity and hinder effective query–table alignment. We propose STAR (Semantic Table Representation), a lightweight framework that improves semantic table representation through semantic clustering and weighted fusion. STAR first applies header-aware K-means clustering to group semantically similar rows and selects representative centroid instances to construct a diverse partial table. It then generates cluster-specific synthetic queries to comprehensively cover the table's semantic space. Finally, STAR employs weighted fusion strategies to integrate table and query embeddings, enabling fine-grained semantic alignment. This design enables STAR to capture complementary information from structured and textual sources, improving the expressiveness of table representations. Experiments on five benchmarks show that STAR achieves consistently higher Recall than QGpT on all datasets, demonstrating the effectiveness of semantic clustering and weighted fusion for robust table representation. Our code is available at https://github.com/adsl135789/STAR.
论文检索
输入标题、作者或关键词，从 954 篇学术成果中精准定位
Foundation models for tabular data, such as the Tabular Prior-data Fitted Network (TabPFN), are pre-trained on a massive number of synthetic datasets generated by structural causal models (SCM). They leverage in-context learning to offer high predictive accuracy in real-world tasks. However, the fairness properties of these foundational models, which incorporate ideas from causal reasoning during pre-training, remain underexplored. In this work, we conduct a comprehensive empirical evaluation of TabPFN and its fine-tuned variants, assessing predictive performance, fairness, and robustness across varying dataset sizes and distributional shifts. Our results reveal that while TabPFN achieves stronger predictive accuracy compared to baselines and exhibits robustness to spurious correlations, improvements in fairness are moderate and inconsistent, particularly under missing-not-at-random (MNAR) covariate shifts. These findings suggest that the causal pre-training in TabPFN is helpful but insufficient for algorithmic fairness, highlighting implications for deploying TabPFN (and similar) models in practice and the need for further fairness interventions.
Academic question answering requires reasoning over heterogeneous scholarly graphs, where queries range from simple attribute lookups to multi-hop inference across author--paper--venue structures. Existing retrieval-augmented generation (RAG) systems struggle in this setting due to three limitations: (1) fixed retrieval strategies that do not adapt to varying query complexity, (2) the absence of sufficiency evaluation leading to incomplete or misaligned evidence, and (3) a lack of structured verification against graph facts. To address these issues, we propose an agentic heterogeneous graph RAG method that transforms the three core stages of the RAG pipeline into explicit agentic decision steps. A query-aware retrieval agent analyzes query type and selects an appropriate graph traversal strategy; a sufficiency-aware reranking agent assesses evidence completeness and adaptively expands the retrieved subgraph; and a graph-grounded verification agent checks entity, relation, and attribute correctness before finalizing the answer. Experiments on heterogeneous graphs constructed from OpenAlex and DBLP suggest that our method consistently outperforms strong LLM, graph-augmented RAG, and agent-based baselines.
We present ReSuMe, a general framework for mutual enhancement of dense retrieval systems and document summarizers through reinforcement learning. The framework jointly optimizes a language model for generating retrieval-oriented summaries and adapts the retrieval model to these summaries through alternating fine-tuning phases. We employ Group Relative Policy Optimization (GRPO) to fine-tune the language model based on retrieval relevance rather than linguistic quality alone, while the retrieval model is iteratively updated using contrastive learning on the generated summaries. This co-optimization process addresses the fundamental distribution shift problem that arises when retrieval models trained on full documents must operate on synthetic summaries during inference. By progressively reducing this distribution gap, our framework yields two key benefits: improved retrieval performance and a high-quality document summarizer optimized for retrieval tasks. We demonstrate our framework using Contriever on the MS-MARCO dataset, achieving consistent improvements of 13.2% in MRR@10 and 6.7% in Recall@100 over the baseline. The framework is model-agnostic and can be applied to enhance any dense retrieval system while simultaneously producing an effective document summarization model.
Large Language Models (LLMs) perform well across many tasks but degrade when processing large collections of repetitive or highly similar inputs, a common scenario in applications such as near-duplicate search results and large e-commerce catalogs. In these settings, concatenation-based approaches—long-context prompting and supervised fine-tuning—suffer from attention saturation and diminished signal-to-noise ratio, causing models to miss subtle but important distinctions as input size grows. We introduce SEER (Set Encoding for Efficient Representation), a framework that enables LLMs to handle massive sets of near-duplicate items through a single learned token. SEER first encodes individual items with a pretrained embedding model, then aggregates them using an adapter that captures inter-item relationships and preserves fine-grained differences while mitigating redundancy. To ensure both discriminative and generative capabilities, we propose a multi-task alignment strategy that supervises set-level descriptions across multiple semantic dimensions. Experiments on a large-scale e-commerce dataset demonstrate that SEER substantially outperforms in-context and fine-tuned LLM baselines, maintaining stable performance even when processing thousands of highly similar items. These results establish SEER as an effective and scalable approach for LLM processing of dense, redundant input sets.
Web searchers continuously form expectations about document content based on snippets and titles, yet when these expectations are violated, their attention and satisfaction are disrupted. Detecting such expectancy violations in real-time can enable adaptive, user-aware Web systems that respond to cognitive mismatches. This paper investigates whether eye-tracking features can predict expectancy violations during Web search and identifies which temporal aspects of attention carry predictive information. Using data from 34 participants performing controlled search tasks, we extracted four gaze metrics, i.e. time to first fixation, total fixation duration, number of fixations, and mean fixation duration, and trained machine learning models using a leave-one-participant-out cross-validation approach. Sustained attention features, particularly total fixation duration and number of fixations, predicted expectancy violations with 62.6% accuracy (p = .008), while initial attention metrics performed at chance. The results reveal that expectancy violations manifest through extended visual engagement rather than immediate orienting responses. The findings of this work contribute to the theoretical understanding of user–system interaction on the Web and provide a foundation for adaptive retrieval interfaces capable of detecting cognitive surprise and delivering timely support.
Topic modeling plays a critical role in organizing and understanding large-scale web content. While neural topic models (NTMs) based on variational autoencoders (VAEs) have achieved notable success in analyzing textual data, they remain limited in addressing the multimodal nature of modern web content. Existing unimodal or multimodal extensions often suffer from posterior collapse and fail to capture the directional semantics inherent in both text and images, resulting in incoherent topics and limited interpretability. To address these challenges, we propose MM-vNTM (MultiModal Neural Topic Model with von Mises-Fisher Mixtures), a framework for web-scale topic discovery over multimodal data. MM-vNTM leverages pre-aligned cross-modal embeddings as inputs and jointly models document-level representations of text and image modalities in a shared hyperspherical latent space. Furthermore, it defines topics as mixtures of von Mises-Fisher (vMF) distributions in the L2-normalized word embedding space, explicitly capturing directional similarity. Experiments on multimedia web datasets demonstrate that MM-vNTM consistently outperforms state-of-the-art unimodal and multimodal baselines in terms of overall topic quality, highlighting its effectiveness for real-world web scenarios.
The rapid proliferation of AI-generated content on the Web presents a structural risk to information retrieval, as search engines and Retrieval-Augmented Generation (RAG) systems increasingly consume evidence produced by the Large Language Models (LLMs). We characterize this ecosystem-level failure mode as Retrieval Collapse, a two-stage process where (1) AI-generated content dominates search results, eroding source diversity, and (2) low-quality or adversarial content infiltrates the retrieval pipeline. We analyzed this dynamic through controlled experiments involving both high-quality SEO-style content and adversarially crafted content. In the SEO scenario, a 67% pool contamination led to over 80% exposure contamination, creating a homogenized yet deceptively healthy state where answer accuracy remains stable despite the reliance on synthetic sources. Conversely, under adversarial contamination, baselines like BM25 exposed ~19% of harmful content, whereas LLM-based rankers demonstrated stronger suppression capabilities. These findings highlight the risk of retrieval pipelines quietly shifting toward synthetic evidence and the need for retrieval-aware strategies to prevent a self-reinforcing cycle of quality decline in Web-grounded systems.
Most graph unlearning evaluations delete a random fraction k% of nodes or edges. In practice, however, controllers act on natural?language erasure requests. To support evaluation under such text inputs, we release Text2Forget, a DSAR-style natural-language erasure-request corpus and benchmark for CoraFull, PubMed, and Yelp, together with a resolver that maps request text to concrete graph targets. Our pipeline designs law?motivated, retrieval-grounded requests that use human-readable anchors and resolves each request to real nodes and edges via a hybrid BM25-plus-embedding index, yielding thousands of requests per dataset. Because this corpus is large, a deterministic ranking score selects a small, actionable top-k subset for evaluation. A plug-and-play harness (GCN, GAT, and GIN with UtU, GNNDelete, GIF, and Retrain) reports privacy via membership-inference AUC (MI-AUC) on forgotten items, averaged over the selected requests and backbones. Across datasets, request-driven forget-sets induce membership-inference and task-accuracy behavior broadly comparable to standard random-forget and retraining baselines, indicating that the Text2Forget corpus can serve as a practical benchmark within existing privacy–utility trade-offs. The released package includes the corpus, resolved targets, confidence scores, seeds, and scripts, supporting reproducible, text-driven unlearning studies. We release our package at https://github.com/ImranAhsan23/Text2Forget.
The performance of large language model (LLM) inference engines heavily relies on deployment-specific parameters such as batch size, parallelism, CUDA graphs, and GPU and network topology, however, with every new LLM released, the optimal setting of these parameters becomes a problem of interest. Additionally, as developers find new applications for LLMs, the parameter space only increases, with large scale inference deployments that span multiple nodes (i.e. llm-d [5]). In this work, we demonstrate ''llm-tuna'', an open-source framework that automates parameter search for vLLM [4] based inference workloads across a diverse set of hardware setups. Our framework utilizes Bayesian optimization via Optuna to select the best configuration while parallelizing trials across multiple nodes to enable large scale studies. We evaluate our system on a mixture-of-experts (MoE) model family (Qwen3-30B-A3B, Kimi-K2-INT4, GPT-OSS-120B, Deepseek-R1-671B), configuring each experiment to avoid artifacts of KV-cache pre-emption effects. Our method achieves output throughput gains of up to 32.9% while never suggesting configurations that are worse than vLLMs' baseline performance. Our results demonstrate the potential of systematic auto-tuning to expose non-obvious performance regimes and guide operator defaults.
Despite significant progress in text anomaly detection for web applications such as spam filtering and fake news detection, existing methods are fundamentally limited to document-level analysis, unable to identify which specific parts of a text are anomalous. We introduce token-level anomaly detection, a novel paradigm that enables fine-grained localization of anomalies within text. We formally define text anomalies at both document and token-levels, and propose a unified detection framework that operates across multiple levels. To facilitate research in this direction, we collect and annotate three benchmark datasets spanning spam, reviews and grammar errors with token-level labels. Experimental results demonstrate that our framework achieves better performance than other 6 baselines, opening new possibilities for precise anomaly localization in text. All the codes and data are publicly available on https://github.com/charles-cao/TokenCore.
Scientific fact-checking is vital for assessing claims in specialized domains such as biomedicine and materials science, yet existing systems often hallucinate or apply inconsistent reasoning, especially when verifying technical, compositional claims against an evidence snippet under source and cost/latency constraints. We present a pipeline centered on atomic predicate–argument decomposition and calibrated, uncertainty-gated corroboration: atomic facts are aligned to local snippets via embeddings, verified by a compact evidence-grounded checker, and only facts with uncertain support trigger domain-restricted web search over authoritative sources. The system supports both binary and tri-valued classification where it predicts labels from Supported, Refuted, NEI for three-way tasks. We evaluate under two regimes, Context-Only (no web) and Context+Web (uncertainty-gated web corroboration); when retrieved evidence conflicts with the provided context, we abstain with NEI rather than overriding the context. On multiple benchmarks, our framework surpasses the strongest benchmarks. In our experiments, web corroboration was invoked for only a minority of atomic facts on average, indicating that external evidence is consulted selectively under calibrated uncertainty rather than routinely. Overall, coupling atomic granularity with calibrated, uncertainty-gated corroboration yields more interpretable and context-conditioned verification, making the approach well-suited to high-stakes, single-document settings that demand traceable rationales, predictable cost/latency, and conservative abstention under domain and evidence grounding constraints.
Trending news detection in low-traffic search environments faces a fundamental cold-start problem, where a lack of query volume prevents systems from identifying emerging or long-tail trends. Existing methods relying on keyword frequency or query spikes are inherently slow and ineffective in these sparse settings, lagging behind real-world shifts in attention. We introduce RTTP, a novel Real-Time Trending Prediction framework that generates search queries directly from news content instead of waiting for users to issue them. RTTP leverages a continual learning LLM (CL-LLM) that converts posts into search-style queries and scores them using engagement strength + creator authority, enabling early trend surfacing before search volume forms. To ensure adaptation without degrading reasoning, we propose Mix-Policy DPO, a new preference-based continual learning approach that combines on-policy stability with off-policy novelty to mitigate catastrophic forgetting during model upgrades. Deployed at production scale on Facebook and Meta AI products, RTTP delivers +91.4% improvement in tail-trend detection precision@500 and +19% query generation accuracy over industry baselines, while sustaining stable performance after multi-week online training. This work demonstrates that LLM-generated synthetic search signals, when aligned and continually updated, unlock timely trend understanding in low-traffic search environments.
The Form 10-K provides a detailed overview of a company's operations, strategies, competitive environment, and future outlook. Investors meticulously analyse annual changes in these reports to identify subtle shifts in disclosed information. However, financial narratives often contain extensive boilerplate text, which dilutes the efficacy of traditional sentiment analysis. We propose a novel edit-level analysis framework to isolate meaningful changes in Management Discussion and Analysis (MD&A) sections of 10-K filings. Using sentence embeddings, we classify each sentence in the filings for years t and t+1 as Kept (repeated boilerplate), Deleted (removed information), and Added (newly introduced information). We then construct a net sentiment change signal from the difference in average FinBERT-tone sentiment between Added and Deleted content. This Core Sentiment measure has been demonstrated to outperform conventional bag-of-words sentiment measures. Furthermore, based on cleaned and standardised 10-K corpora (e.g., EDGAR-CORPUS) and our Flex10K pre-processing pipeline, our approach provides a reproducible basis for quantifying changes between annual reports.
Concerns about AI-generated political content are growing, yet there is limited empirical evidence on how deepfakes appear and circulate across social platforms during major events in democratic countries. We analyze the 2025 Canadian federal election across X, Bluesky, and Reddit using a high-accuracy detector trained on diverse modern generative models, covering 187,778 posts. We find that 5.9% of election-related images were deepfakes. Right-leaning accounts shared them more often (9.2% of images flagged) than left-leaning users (3.9%), with flagged content more frequently defamatory or conspiratorial. Yet, most detected deepfakes were benign or non-political, and harmful ones drew little attention, accounting for only 0.1% of all views on X. Overall, deepfakes were present in the election conversation, but their reach was modest, and realistic fabricated images, although less common, drew higher engagement, highlighting growing concerns about their misuses.
Fraudulent activities on blockchain networks threaten the integrity and reliability of decentralized finance ecosystems. Accurately identifying malicious nodes such as phishing or ransomware addresses, within large-scale blockchain transaction graphs remains a critical challenge due to their dynamic, sparse, and continuously evolving topologies. Transfer learning offers a powerful paradigm for fraud detection because many fraudulent schemes, including ransomware and phishing, are often orchestrated by overlapping actor groups that share behavioral and structural patterns across networks. Leveraging these shared representations enables knowledge transfer from previously observed fraud types to emerging ones. However, the complex and multi-modal nature of digital financial systems introduces substantial challenges for graph-based transfer learning. Fraudulent activities are shaped by diverse modalities including graph structure, transaction sequences, temporal price dynamics, and textual metadata, while distributional shifts frequently occur across time and platforms. Existing graph transfer learning methods struggle to model such multi-modal dependencies and to align divergent feature distributions. To tackle these challenges, we develop a Multi-mOdal Enhanced Graph Transfer Learning (MOE-GTL) framework which incorporates graph, temporal, and textual modalities for fraudulent node detection. We further introduce Temporal-aware Maximum Mean Discrepancy (TMMD), a regularization mechanism that explicitly aligns multi-modal feature distributions between source and target graphs over time. Extensive experiments reveal that our MOE-GTL model notably improves the accuracy of fraudulent node classifications on Ethereum and Solana transaction graphs.
Municipal meeting minutes are formal records documenting the discussions and decisions of local government, yet their content is often lengthy, dense, and difficult for citizens to navigate. Automatic summarization can help address this challenge by producing concise summaries for each discussion subject. Despite its potential, research on summarizing discussion subjects in municipal meeting minutes remains largely unexplored, especially in low-resource languages, where the inherent complexity of these documents adds further challenges. A major bottleneck is the scarcity of datasets containing high-quality, manually crafted summaries, which limits the development and evaluation of effective summarization models for this domain. In this paper, we present CitiLink-Summ, a new corpus of European Portuguese municipal meeting minutes, comprising 120 documents and 2,880 manually hand-written summaries, each corresponding to a distinct discussion subject. Leveraging this dataset, we establish baseline results for automatic summarization in this domain, employing state-of-the-art generative models (e.g., BART, PRIMERA) as well as large language models (LLMs), evaluated with both lexical and semantic metrics such as ROUGE, BLEU, METEOR, and BERTScore. CitiLink-Summ provides the first benchmark for municipal-domain summarization in European Portuguese, offering a valuable resource for advancing NLP research on complex administrative texts.
In this paper, we propose Pr ecision-Informed Semantic Modeling (PRISM), a structured topic modeling framework combining the benefits of rich representations captured by LLMs with the low cost and interpretability of latent semantic clustering methods. PRISM fine-tunes a sentence encoding model using a sparse set of LLM-provided labels on samples drawn from some corpus of interest. We segment this embedding space with thresholded clustering, yielding clusters that separate closely related topics within some narrow domain. Across multiple corpora, PRISM improves topic separability over state-of-the-art local topic models and even over clustering on large, frontier embedding models while requiring only a small number of LLM queries to train. This work contributes to several research streams by providing (i) a student-teacher pipeline to distill sparse LLM supervision into a lightweight model for topic discovery; (ii) an analysis of the efficacy of sampling strategies to improve local geometry for cluster separability; and (iii)~an effective approach for web-scale text analysis, enabling researchers and practitioners to track nuanced claims and subtopics online with an interpretable, locally deployable framework.
Online marketplaces will be transformed by autonomous AI agents acting on behalf of consumers. Rather than humans browsing and clicking, AI agents can parse webpages or interact through APIs to evaluate products, and transact. This raises a fundamental question: what do AI agents buy—and why? We develop ACES, a sandbox environment that pairs a platform-agnostic agent with a fully programmable mock marketplace to study this. We first explore aggregate choices, revealing that modal choices can differ across models, with AI agents sometimes concentrating on a few products, raising competition questions. We then analyze the current drivers of choices through randomized experiments on product positions and listing attributes. Models show sizeable and heterogeneous position effects: all favor the top row, yet different models prefer different columns, undermining the assumption of a universal ''top'' rank. They penalize sponsored tags, reward endorsements, and sensitivities to price, ratings, and reviews are directionally as expected, but vary sharply across models. Our findings reveal how AI agents behave in e-commerce, and surface concrete monitoring, seller strategy, platform design, and regulatory questions.
This study investigates the problem of keeping information up-to-date when crawling data sources that change over time (e.g., websites or location data). Traditional crawling methods often treat data sources independently, making it difficult to capture relationships and propagate updates efficiently. We propose using graph structures to model these relationships and show that, unfortunately, finding the theoretically optimal solution can be intractable. To address this, we introduce a specific graphical model (the latent Bernoulli process model) and demonstrate the complexity of even simple tasks within this framework. We tackle the crawling problem using a reinforcement learning-based algorithm and demonstrate its superiority over traditional baselines on both real and synthetic data. This work highlights the power of graph-structured crawling in helping users stay informed within a dynamic information landscape.