arXiv subjects

Xiaoyang Wang

Publications and source records attributed to Xiaoyang Wang.

At least 109 records · Page 6Linked to original sources

arXiv2025

DSBench: How Far Are Data Science Agents from Becoming Data Science Experts?

Large Language Models (LLMs) and Large Vision-Language Models (LVLMs) have demonstrated impressive language/vision reasoning abilities, igniting the recent trend of building agents for targeted applications such as shopping assistants or AI software engineers. Recently, many data science benchmarks have been proposed to investigate their performance in the data science domain. However, existing data science benchmarks still fall short when compared to real-world data science applications due to their simplified settings. To bridge this gap, we introduce DSBench, a comprehensive benchmark designed to evaluate data science agents with realistic tasks. This benchmark includes 466 data analysis tasks and 74 data modeling tasks, sourced from Eloquence and Kaggle competitions. DSBench offers a realistic setting by encompassing long contexts, multimodal task backgrounds, reasoning with large data files and multi-table structures, and performing end-to-end data modeling tasks. Our evaluation of state-of-the-art LLMs, LVLMs, and agents shows that they struggle with most tasks, with the best agent solving only 34.12% of data analysis tasks and achieving a 34.74% Relative Performance Gap (RPG). These findings underscore the need for further advancements in developing more practical, intelligent, and autonomous data science agents.

cs.AI↗

arXiv2025

Paths-over-Graph: Knowledge Graph Empowered Large Language Model Reasoning

Large Language Models (LLMs) have achieved impressive results in various tasks but struggle with hallucination problems and lack of relevant knowledge, especially in deep complex reasoning and knowledge-intensive tasks. Knowledge Graphs (KGs), which capture vast amounts of facts in a structured format, offer a reliable source of knowledge for reasoning. However, existing KG-based LLM reasoning methods face challenges like handling multi-hop reasoning, multi-entity questions, and effectively utilizing graph structures. To address these issues, we propose Paths-over-Graph (PoG), a novel method that enhances LLM reasoning by integrating knowledge reasoning paths from KGs, improving the interpretability and faithfulness of LLM outputs. PoG tackles multi-hop and multi-entity questions through a three-phase dynamic multi-hop path exploration, which combines the inherent knowledge of LLMs with factual knowledge from KGs. In order to improve the efficiency, PoG prunes irrelevant information from the graph exploration first and introduces efficient three-step pruning techniques that incorporate graph structures, LLM prompting, and a pre-trained language model (e.g., SBERT) to effectively narrow down the explored candidate paths. This ensures all reasoning paths contain highly relevant information captured from KGs, making the reasoning faithful and interpretable in problem-solving. PoG innovatively utilizes graph structure to prune the irrelevant noise and represents the first method to implement multi-entity deep path detection on KGs for LLM reasoning tasks. Comprehensive experiments on five benchmark KGQA datasets demonstrate PoG outperforms the state-of-the-art method ToG across GPT-3.5-Turbo and GPT-4, achieving an average accuracy improvement of 18.9%. Notably, PoG with GPT-3.5-Turbo surpasses ToG with GPT-4 by up to 23.9%.

cs.CL↗

arXiv2025

Discovery of High-Temperature Superconducting Ternary Hydrides via Deep Learning

The discovery of novel high-temperature superconductor materials holds transformative potential for a wide array of technological applications. However, the combinatorially vast chemical and configurational search space poses a significant bottleneck for both experimental and theoretical investigations. In this study, we employ the design of high-temperature ternary superhydride superconductors as a representative case to demonstrate how this challenge can be well addressed through a deep-learning-driven theoretical framework. This framework integrates high-throughput crystal structure exploration, physics-informed screening, and accurate prediction of superconducting critical temperatures. Our approach enabled the exploration of approximately 36 million ternary hydride structures across a chemical space of 29 elements, leading to the identification of 144 potential high-Tc superconductors with predicted Tc > 200 K and superior thermodynamic stability at 200 GPa. Among these, 129 compounds spanning 27 novel structural prototypes are reported for the first time, representing a significant expansion of the known structural landscape for hydride superconductors. This work not only greatly expands the known repertoire of high-Tc hydride superconductors but also establishes a scalable and efficient methodology for navigating the complex landscape of multinary hydrides.

cond-mat.supr-con↗

arXiv2025

OpenCharacter: Training Customizable Role-Playing LLMs with Large-Scale Synthetic Personas

Customizable role-playing in large language models (LLMs), also known as character generalization, is gaining increasing attention for its versatility and cost-efficiency in developing and deploying role-playing dialogue agents. This study explores a large-scale data synthesis approach to equip LLMs with character generalization capabilities. We begin by synthesizing large-scale character profiles using personas from Persona Hub and then explore two strategies: response rewriting and response generation, to create character-aligned instructional responses. To validate the effectiveness of our synthetic instruction tuning data for character generalization, we perform supervised fine-tuning (SFT) using the LLaMA-3 8B model. Our best-performing model strengthens the original LLaMA-3 8B Instruct model and achieves performance comparable to GPT-4o models on role-playing dialogue. We release our synthetic characters and instruction-tuning dialogues to support public research.

cs.CL↗

arXiv2025

RIDA: A Robust Attack Framework on Incomplete Graphs

Graph Neural Networks (GNNs) are vital in data science but are increasingly susceptible to adversarial attacks. To help researchers develop more robust GNN models, it's essential to focus on designing strong attack models as foundational benchmarks and guiding references. Among adversarial attacks, gray-box poisoning attacks are noteworthy due to their effectiveness and fewer constraints. These attacks exploit GNNs' need for retraining on updated data, thereby impacting their performance by perturbing these datasets. However, current research overlooks the real-world scenario of incomplete graphs. To address this gap, we introduce the Robust Incomplete Deep Attack Framework (RIDA). It is the first algorithm for robust gray-box poisoning attacks on incomplete graphs. The approach innovatively aggregates distant vertex information and ensures powerful data utilization. Extensive tests against 9 SOTA baselines on 3 real-world datasets demonstrate that RIDA's superiority in handling incompleteness and high attack performance on the incomplete graph.

cs.LG↗

arXiv2025

Enhancing Multi-Attribute Fairness in Healthcare Predictive Modeling

Artificial intelligence (AI) systems in healthcare have demonstrated remarkable potential to improve patient outcomes. However, if not designed with fairness in mind, they also carry the risks of perpetuating or exacerbating existing health disparities. Although numerous fairness-enhancing techniques have been proposed, most focus on a single sensitive attribute and neglect the broader impact that optimizing fairness for one attribute may have on the fairness of other sensitive attributes. In this work, we introduce a novel approach to multi-attribute fairness optimization in healthcare AI, tackling fairness concerns across multiple demographic attributes concurrently. Our method follows a two-phase approach: initially optimizing for predictive performance, followed by fine-tuning to achieve fairness across multiple sensitive attributes. We develop our proposed method using two strategies, sequential and simultaneous. Our results show a significant reduction in Equalized Odds Disparity (EOD) for multiple attributes, while maintaining high predictive accuracy. Notably, we demonstrate that single-attribute fairness methods can inadvertently increase disparities in non-targeted attributes whereas simultaneous multi-attribute optimization achieves more balanced fairness improvements across all attributes. These findings highlight the importance of comprehensive fairness strategies in healthcare AI and offer promising directions for future research in this critical area.

cs.LG↗

arXiv2025

Addressing the sustainable AI trilemma: a case study on LLM agents and RAG

Large language models (LLMs) have demonstrated significant capabilities, but their widespread deployment and more advanced applications raise critical sustainability challenges, particularly in inference energy consumption. We propose the concept of the Sustainable AI Trilemma, highlighting the tensions between AI capability, digital equity, and environmental sustainability. Through a systematic case study of LLM agents and retrieval-augmented generation (RAG), we analyze the energy costs embedded in memory module designs and introduce novel metrics to quantify the trade-offs between energy consumption and system performance. Our experimental results reveal significant energy inefficiencies in current memory-augmented frameworks and demonstrate that resource-constrained environments face disproportionate efficiency penalties. Our findings challenge the prevailing LLM-centric paradigm in agent design and provide practical insights for developing more sustainable AI systems.

cs.CY↗

arXiv2025

Deep Overlapping Community Search via Subspace Embedding

Overlapping Community Search (OCS) identifies nodes that interact with multiple communities based on a specified query. Existing community search approaches fall into two categories: algorithm-based models and Machine Learning-based (ML) models. Despite the long-standing focus on this topic within the database domain, current solutions face two major limitations: 1) Both approaches fail to address personalized user requirements in OCS, consistently returning the same set of nodes for a given query regardless of user differences. 2) Existing ML-based CS models suffer from severe training efficiency issues. In this paper, we formally redefine the problem of OCS. By analyzing the gaps in both types of approaches, we then propose a general solution for OCS named Sparse Subspace Filter (SSF), which can extend any ML-based CS model to enable personalized search in overlapping structures. To overcome the efficiency issue in the current models, we introduce Simplified Multi-hop Attention Networks (SMN), a lightweight yet effective community search model with larger receptive fields. To the best of our knowledge, this is the first ML-based study of overlapping community search. Extensive experiments validate the superior performance of SMN within the SSF pipeline, achieving a 13.73% improvement in F1-Score and up to 3 orders of magnitude acceleration in model efficiency compared to state-of-the-art approaches.

cs.SI↗

arXiv2025

Imaginary Hamiltonian variational ansatz for combinatorial optimization problems

Obtaining exact solutions to combinatorial optimization problems using classical computing is computationally expensive. The current tenet in the field is that quantum computers can address these problems more efficiently. While promising algorithms require fault-tolerant quantum hardware, variational algorithms have emerged as viable candidates for near-term devices. The success of these algorithms hinges on multiple factors, with the design of the ansatz having the utmost importance. It is known that popular approaches such as quantum approximate optimization algorithm (QAOA) and quantum annealing suffer from adiabatic bottlenecks, that lead to either larger circuit depth or evolution time. On the other hand, the evolution time of imaginary time evolution is bounded by the inverse energy gap of the Hamiltonian, which is constant for most non-critical physical systems. In this work, we propose imaginary Hamiltonian variational ansatz ($i$HVA) inspired by quantum imaginary time evolution to solve the MaxCut problem. We introduce a tree arrangement of the parametrized quantum gates, enabling the exact solution of arbitrary tree graphs using the one-round $i$HVA. For randomly generated $D$-regular graphs, we numerically demonstrate that the $i$HVA solves the MaxCut problem with a small constant number of rounds and sublinear depth, outperforming QAOA, which requires rounds increasing with the graph size. Furthermore, our ansatz solves MaxCut exactly for graphs with up to 24 nodes and $D \leq 5$, whereas only approximate solutions can be derived by the classical near-optimal Goemans-Williamson algorithm. We validate our simulated results with hardware demonstrations on a graph with 67 nodes.

quant-ph↗

arXiv2024

CNC: Cross-modal Normality Constraint for Unsupervised Multi-class Anomaly Detection

Existing unsupervised distillation-based methods rely on the differences between encoded and decoded features to locate abnormal regions in test images. However, the decoder trained only on normal samples still reconstructs abnormal patch features well, degrading performance. This issue is particularly pronounced in unsupervised multi-class anomaly detection tasks. We attribute this behavior to over-generalization(OG) of decoder: the significantly increasing diversity of patch patterns in multi-class training enhances the model generalization on normal patches, but also inadvertently broadens its generalization to abnormal patches. To mitigate OG, we propose a novel approach that leverages class-agnostic learnable prompts to capture common textual normality across various visual patterns, and then apply them to guide the decoded features towards a normal textual representation, suppressing over-generalization of the decoder on abnormal patterns. To further improve performance, we also introduce a gated mixture-of-experts module to specialize in handling diverse patch patterns and reduce mutual interference between them in multi-class training. Our method achieves competitive performance on the MVTec AD and VisA datasets, demonstrating its effectiveness.

cs.CV↗

arXiv2024

Efficient Dynamic Attributed Graph Generation

Data generation is a fundamental research problem in data management due to its diverse use cases, ranging from testing database engines to data-specific applications. However, real-world entities often involve complex interactions that cannot be effectively modeled by traditional tabular data. Therefore, graph data generation has attracted increasing attention recently. Although various graph generators have been proposed in the literature, there are three limitations: i) They cannot capture the co-evolution pattern of graph structure and node attributes. ii) Few of them consider edge direction, leading to substantial information loss. iii) Current state-of-the-art dynamic graph generators are based on the temporal random walk, making the simulation process time-consuming. To fill the research gap, we introduce VRDAG, a novel variational recurrent framework for efficient dynamic attributed graph generation. Specifically, we design a bidirectional message-passing mechanism to encode both directed structural knowledge and attribute information of a snapshot. Then, the temporal dependency in the graph sequence is captured by a recurrence state updater, generating embeddings that can preserve the evolution pattern of early graphs. Based on the hidden node embeddings, a conditional variational Bayesian method is developed to sample latent random variables at the neighboring timestep for new snapshot generation. The proposed generation paradigm avoids the time-consuming path sampling and merging process in existing random walk-based methods, significantly reducing the synthesis time. Finally, comprehensive experiments on real-world datasets are conducted to demonstrate the effectiveness and efficiency of the proposed model.

cs.DB↗

arXiv2024

Parallel Higher-order Truss Decomposition

The k-truss model is one of the most important models in cohesive subgraph analysis. The k-truss decomposition problem is to compute the trussness of each edge in a given graph, and has been extensively studied. However, the conventional k-truss model is difficult to characterize the fine-grained hierarchical structures in networks due to the neglect of high order information. To overcome the limitation, the higher-order truss model is proposed in the literature. However, the previous solutions only consider non-parallel scenarios. To fill the gap, in this paper, we conduct the first research to study the problem of parallel higher-order truss decomposition. Specifically, a parallel framework is first proposed. Moreover, several optimizations are further developed to accelerate the processing. Finally, experiments over 6 real-world networks are conducted to verify the performance of proposed methods.

cs.DS↗

arXiv2024

NaturalConv: A Chinese Dialogue Dataset Towards Multi-turn Topic-driven Conversation

In this paper, we propose a Chinese multi-turn topic-driven conversation dataset, NaturalConv, which allows the participants to chat anything they want as long as any element from the topic is mentioned and the topic shift is smooth. Our corpus contains 19.9K conversations from six domains, and 400K utterances with an average turn number of 20.1. These conversations contain in-depth discussions on related topics or widely natural transition between multiple topics. We believe either way is normal for human conversation. To facilitate the research on this corpus, we provide results of several benchmark models. Comparative results show that for this dataset, our current models are not able to provide significant improvement by introducing background knowledge/topic. Therefore, the proposed dataset should be a good benchmark for further research to evaluate the validity and naturalness of multi-turn conversation systems. Our dataset is available at https://ailab.tencent.com/ailab/nlp/dialogue/#datasets.

cs.CL↗

arXiv2024

TCGU: Data-centric Graph Unlearning based on Transferable Condensation

With growing demands for data privacy and model robustness, graph unlearning (GU), which erases the influence of specific data on trained GNN models, has gained significant attention. However, existing exact unlearning methods suffer from either low efficiency or poor model performance. While being more utility-preserving and efficient, current approximate unlearning methods are not applicable in the zero-glance privacy setting, where the deleted samples cannot be accessed during unlearning due to immediate deletion requested by regulations. Besides, these approximate methods, which try to directly perturb model parameters still involve high privacy concerns in practice. To fill the gap, we propose Transferable Condensation Graph Unlearning (TCGU), a data-centric solution to zero-glance graph unlearning. Specifically, we first design a two-level alignment strategy to pre-condense the original graph into a small yet utility-preserving dataset. Upon receiving an unlearning request, we fine-tune the pre-condensed data with a low-rank plugin, to directly align its distribution with the remaining graph, thus efficiently revoking the information of deleted data without accessing them. A novel similarity distribution matching approach and a discrimination regularizer are proposed to effectively transfer condensed data and preserve its utility in GNN training, respectively. Finally, we retrain the GNN on the transferred condensed data. Extensive experiments on 6 benchmark datasets demonstrate that TCGU can achieve superior performance in terms of model utility, unlearning efficiency, and unlearning efficacy than existing GU methods.

cs.LG↗

arXiv2024

When Reasoning Meets Information Aggregation: A Case Study with Sports Narratives

Reasoning is most powerful when an LLM accurately aggregates relevant information. We examine the critical role of information aggregation in reasoning by requiring the LLM to analyze sports narratives. To succeed at this task, an LLM must infer points from actions, identify related entities, attribute points accurately to players and teams, and compile key statistics to draw conclusions. We conduct comprehensive experiments with real NBA basketball data and present SportsGen, a new method to synthesize game narratives. By synthesizing data, we can rigorously evaluate LLMs' reasoning capabilities under complex scenarios with varying narrative lengths and density of information. Our findings show that most models, including GPT-4o, often fail to accurately aggregate basketball scores due to frequent scoring patterns. Open-source models like Llama-3 further suffer from significant score hallucinations. Finally, the effectiveness of reasoning is influenced by narrative complexity, information density, and domain-specific terms, highlighting the challenges in analytical reasoning tasks.

cs.CL↗

arXiv2024

FakeBench: Probing Explainable Fake Image Detection via Large Multimodal Models

The ability to distinguish whether an image is generated by artificial intelligence (AI) is a crucial ingredient in human intelligence, usually accompanied by a complex and dialectical forensic and reasoning process. However, current fake image detection models and databases focus on binary classification without understandable explanations for the general populace. This weakens the credibility of authenticity judgment and may conceal potential model biases. Meanwhile, large multimodal models (LMMs) have exhibited immense visual-text capabilities on various tasks, bringing the potential for explainable fake image detection. Therefore, we pioneer the probe of LMMs for explainable fake image detection by presenting a multimodal database encompassing textual authenticity descriptions, the FakeBench. For construction, we first introduce a fine-grained taxonomy of generative visual forgery concerning human perception, based on which we collect forgery descriptions in human natural language with a human-in-the-loop strategy. FakeBench examines LMMs with four evaluation criteria: detection, reasoning, interpretation and fine-grained forgery analysis, to obtain deeper insights into image authenticity-relevant capabilities. Experiments on various LMMs confirm their merits and demerits in different aspects of fake image detection tasks. This research presents a paradigm shift towards transparency for the fake image detection area and reveals the need for greater emphasis on forensic elements in visual-language research and AI risk control. FakeBench will be available at https://github.com/Yixuan423/FakeBench.

cs.CV↗

arXiv2024

Using Process Mining to Improve Digital Service Delivery

We present a case study of Process Mining (PM) for personnel security screening in the Canadian government. We consider customer (process time) and organizational (cost) perspectives. Furthermore, in contrast to most published case studies, we assess the full process improvement lifecycle: pre-intervention analyses pointed out initial bottlenecks, and post-intervention analyses identified the intervention impact and remaining areas for improvement. Using PM techniques, we identified frequent exceptional scenarios (e.g., applications requiring amendment), time-intensive loops (e.g., employees forgetting tasks), and resource allocation issues (e.g., involvement of non-security personnel). Subsequent process improvement interventions, implemented using a flexible low-code digital platform, reduced security briefing times from around 7 days to 46 hours, and overall process time from around 31 days to 26 days, on average. From a cost perspective, the involvement of hiring managers and security screening officers was significantly reduced. These results demonstrate how PM can become part of a broader digital transformation framework to improve public service delivery. The success of these interventions motivated subsequent government PM projects, and inspired a PM methodology, currently under development, for use in large organizational contexts such as governments.

cs.CY↗

arXiv2024

Coordinate-Aware Thermal Infrared Tracking Via Natural Language Modeling

Thermal infrared (TIR) tracking is pivotal in computer vision tasks due to its all-weather imaging capability. Traditional tracking methods predominantly rely on hand-crafted features, and while deep learning has introduced correlation filtering techniques, these are often constrained by rudimentary correlation operations. Furthermore, transformer-based approaches tend to overlook temporal and coordinate information, which is critical for TIR tracking that lacks texture and color information. In this paper, to address these issues, we apply natural language modeling to TIR tracking and propose a coordinate-aware thermal infrared tracking model called NLMTrack, which enhances the utilization of coordinate and temporal information. NLMTrack applies an encoder that unifies feature extraction and feature fusion, which simplifies the TIR tracking pipeline. To address the challenge of low detail and low contrast in TIR images, on the one hand, we design a multi-level progressive fusion module that enhances the semantic representation and incorporates multi-scale features. On the other hand, the decoder combines the TIR features and the coordinate sequence features using a causal transformer to generate the target sequence step by step. Moreover, we explore an adaptive loss aimed at elevating tracking accuracy and a simple template update strategy to accommodate the target's appearance variations. Experiments show that NLMTrack achieves state-of-the-art performance on multiple benchmarks. The Code is publicly available at \url{https://github.com/ELOESZHANG/NLMTrack}.

cs.CV↗