Information Processing & Management最新文献_第10页

Prompt robust large language model for Chinese medical named entity recognition

IF 7.4 1区管理学

Information Processing & Management Pub Date : 2025-05-12 DOI: 10.1016/j.ipm.2025.104189

Yubo Chen , Baoli Zhang , Sirui Li , Zhuoran Jin , Zhengyuan Cai , Yingzheng Wang , Delai Qiu , ShengPing Liu , Jun Zhao

{"title":"Prompt robust large language model for Chinese medical named entity recognition","authors":"Yubo Chen , Baoli Zhang , Sirui Li , Zhuoran Jin , Zhengyuan Cai , Yingzheng Wang , Delai Qiu , ShengPing Liu , Jun Zhao","doi":"10.1016/j.ipm.2025.104189","DOIUrl":"10.1016/j.ipm.2025.104189","url":null,"abstract":"<div><div>Medical Named Entity Recognition (NER) is crucial for constructing healthcare knowledge graphs and enhancing intelligent medical systems, yet it faces three challenges: data scarcity, low recall in nested entities annotation and high prompt sensitivity of generative NER model. In this paper, we aim to address the three challenges simultaneously. First, we construct a Multi-Scenario Medical NER dataset which is the largest medical NER dataset, including over 40,000 samples and over 3400 entity types with eight major scenarios: medical web data, online consultation, medical book, etc. Second, we propose a decomposed question answering based data annotation and selection method, which improved F1 score by 6% compared to direct annotation. Third, to enhance the robustness of large models to diverse prompts in real-world scenarios, we construct diverse prompt templates and implements dynamic prompt strategy during the training phase. Finally, we conducted a comprehensive set of experiments, and the results demonstrate the effectiveness of our annotation method and robustness training approach. Notably, the proposed framework achieves a 5% performance improvement on the test set compared to conventional methods. Moreover, our method enables a 7B parameter model to surpass a 32B parameter model, highlighting its superior efficiency and capability.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"62 5","pages":"Article 104189"},"PeriodicalIF":7.4,"publicationDate":"2025-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143943412","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Triplet-modality group-guided incremental distillation with regularized group semantic consistency for multi-modal neural machine translation

IF 7.4 1区管理学

Information Processing & Management Pub Date : 2025-05-12 DOI: 10.1016/j.ipm.2025.104149

Junjun Guo, Yunyue Li, Kaiwen Tan

{"title":"Triplet-modality group-guided incremental distillation with regularized group semantic consistency for multi-modal neural machine translation","authors":"Junjun Guo, Yunyue Li, Kaiwen Tan","doi":"10.1016/j.ipm.2025.104149","DOIUrl":"10.1016/j.ipm.2025.104149","url":null,"abstract":"<div><div>Multi-modal Machine Translation (MMT) aims to tackle the challenge of cross-modal semantic alignment by integrating additional information from additional modalities, such as images, video, audio, and potentially other modalities. Unfortunately, collecting high-quality multi-modal data pairs is costly, leading to challenges in data scarcity or noise robustness. Most existing MMT research focuses on feature-level cross-modal fusion using these limited multi-modal data, training models from scratch without utilizing prior knowledge from established pure-text neural machine translation (NMT) models. This results in inefficient use of computational resources and cross-modal misalignment. To this end, this paper presents a triplet-modality group-guided incremental distillation approach, constrained by group-centered multi-modal semantic alignment, to extend the scope of machine translation in visual scenarios. The proposed approach preserves the translation capabilities of the pre-trained NMT model through triplet-modal group incremental distillation, while further improving translation performance through a regularized group alignment strategy, thereby enhancing machine translation ability in MMT. We conducted extensive experiments on two general-domain and two specific-domain MMT tasks. The results demonstrate that the proposed approach shows improvements over the state-of-the-art (SOTA) methods across all test sets, achieving performance gains of over 3.7%. In-depth analysis highlights the effectiveness and robustness of our method in cross-modal alignment and noisy visual scenarios. Our code is available at <span><span>https://github.com/lyy-nlp/MMT_main</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"62 5","pages":"Article 104149"},"PeriodicalIF":7.4,"publicationDate":"2025-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143943413","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Natural vs programming language in LLM knowledge graph construction 法学硕士知识图谱构建中的自然语言与编程语言

IF 7.4 1区管理学

Information Processing & Management Pub Date : 2025-05-09 DOI: 10.1016/j.ipm.2025.104195

Paolo Gajo, Alberto Barrón-Cedeño

{"title":"Natural vs programming language in LLM knowledge graph construction","authors":"Paolo Gajo, Alberto Barrón-Cedeño","doi":"10.1016/j.ipm.2025.104195","DOIUrl":"10.1016/j.ipm.2025.104195","url":null,"abstract":"<div><div>Research on knowledge graph construction (KGC) has recently shown great promise also thanks to the adoption of large language models (LLM) for the automatic extraction of structured information from raw text. However, most works rely on commercial, closed-source LLMs, hindering reproducibility and accessibility. We explore KGC with smaller, open-weight LLMs and investigate whether they can be used to improve upon the results obtained by systems leveraging bigger, closed-source models. Specifically, we focus on CodeKGC, a prompting framework based on GPT-3.5. We choose a variety of models either pre-trained primarily on natural language or on code and fine-tune them on three datasets used for information extraction. We fine-tune with prompts formatted either in natural language or as Python-like scripts. In addition, we optionally train the models with prompts including chain-of-thought sections. After fine-tuning, the choice of coding vs natural language prompts has a limited impact on performance, while chain-of-thought training mostly leads to a performance decrease. Moreover, we show that a LLM can be outperformed by much smaller versions on this task, after undergoing the same amount of training. We find that in general the selected lightweight LLMs outperform the much larger CodeKGC by as much as 15–20 absolute F<span><math><msub><mrow></mrow><mrow><mn>1</mn></mrow></msub></math></span> points after fine-tuning. The results show that state-of-the-art KGC systems can be developed using smaller and open-weight models, enhancing research transparency, lowering compute requirements, and decreasing third-party API reliance.</div><div>Code:</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"62 5","pages":"Article 104195"},"PeriodicalIF":7.4,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143927475","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Structured knowledge-based causal discovery: Agentic streams of thought 结构化的基于知识的因果发现：思想的代理流

IF 7.4 1区管理学

Information Processing & Management Pub Date : 2025-05-09 DOI: 10.1016/j.ipm.2025.104202

Sven Meier, Pratik Narendra Raut, Felix Mahr, Nils Thielen, Jörg Franke, Florian Risch

{"title":"Structured knowledge-based causal discovery: Agentic streams of thought","authors":"Sven Meier, Pratik Narendra Raut, Felix Mahr, Nils Thielen, Jörg Franke, Florian Risch","doi":"10.1016/j.ipm.2025.104202","DOIUrl":"10.1016/j.ipm.2025.104202","url":null,"abstract":"<div><div>Causal discovery—the systematic identification of cause-and-effect relationships among variables—forms the cornerstone of causal inference. Its application enables reliable predictions and targeted interventions across complex systems, from medical treatments to engineering processes. Traditional statistical causal discovery methods face significant limitations with high-dimensional data structures, while existing knowledge-based approaches rely on single large-scale models that raise fundamental concerns about computational efficiency and result reliability. The Agentic Stream of Thought (ASoT) addresses these limitations through a novel architecture that orchestrates multiple smaller open-source language models. The framework integrates hierarchical query decomposition with Model Compiler refinement, while dual-stream thought processing enables balanced analysis through parallel evaluation of competing hypotheses. Dedicated Direction and Transitive Processors enhance reasoning by resolving bidirectional relationships and refining transitive pathways. A two-tiered quality gate system and complementary consensus mechanisms—Delphi protocol and Ensemble Synthesis Method—iteratively refine outputs while mitigating hallucination risks. Empirical evaluations across causal discovery benchmarks and question-answering tasks demonstrate that this approach matches or exceeds state-of-the-art models while enabling local deployment, establishing that sophisticated orchestration of smaller models provides a more sustainable path than increasing model scale alone.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"62 5","pages":"Article 104202"},"PeriodicalIF":7.4,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143927474","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Bi-directional feature learning-based approach for zero-shot event argument extraction 基于双向特征学习的零射击事件参数提取方法

IF 7.4 1区管理学

Information Processing & Management Pub Date : 2025-05-03 DOI: 10.1016/j.ipm.2025.104199

Ji Zhou , Kai Shuang , Qiwei Wang , Bing Qian , Jinyu Guo

{"title":"Bi-directional feature learning-based approach for zero-shot event argument extraction","authors":"Ji Zhou , Kai Shuang , Qiwei Wang , Bing Qian , Jinyu Guo","doi":"10.1016/j.ipm.2025.104199","DOIUrl":"10.1016/j.ipm.2025.104199","url":null,"abstract":"<div><div>Recent research has shown that event argument extraction (EAE) methods based on transfer learning and data augmentation emphasize the contribution of contextual features and labeled features to zero-shot EAE tasks, respectively. However, these methods suffer from knowledge transfer insufficiency and context generation bias challenges. In this paper, we propose a bi-directional feature learning-based approach for zero-shot event argument extraction (BiTer), which gains bi-directional transferable knowledge and mitigates context generation bias. Specifically, BiTer contains source and target model training. During source model training, BiTer co-trains the contextual and labeled feature learning tasks on the source dataset. This step enables the target model to acquire bi-directional transferable knowledge, providing more appropriate feature representations for target events. In target model training, BiTer leverages the large language model to produce pseudo-arguments, and then the knowledge-embedded model generates training data of the target events based on them. This step mitigates context generation bias and makes BiTer learn a more comprehensive and precise feature of the target event. Extensive experiments on RAMS, WIKIEVENTS and ACE2005 have demonstrated BiTer achieves a new state-of-the-art level, with F1 in the zero-shot setting outperforming the baseline model by 4.6%, 7.5% and 0.4%, respectively.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"62 5","pages":"Article 104199"},"PeriodicalIF":7.4,"publicationDate":"2025-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143899218","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Ordinal focal loss: A relative fashionability ranking learning method 有序失焦：一种相对时尚的排序学习方法

IF 7.4 1区管理学

Information Processing & Management Pub Date : 2025-05-03 DOI: 10.1016/j.ipm.2025.104205

Ling Ma , Chuhang Zou , Ziyi Guo , Tao Li , Zheng Liu , Fengyuan Zou

{"title":"Ordinal focal loss: A relative fashionability ranking learning method","authors":"Ling Ma , Chuhang Zou , Ziyi Guo , Tao Li , Zheng Liu , Fengyuan Zou","doi":"10.1016/j.ipm.2025.104205","DOIUrl":"10.1016/j.ipm.2025.104205","url":null,"abstract":"<div><div>Existing fashion recommendations often rely on natural language processing or content-based image retrieval, overlooking direct aesthetic assessments of fashion images. Given the subjectivity and complexity of this task, we propose treating fashionability as a relative attribute to rank paired clothing images. To address this ranking challenge, we propose Ordinal Focal Loss, which transforms the pairwise ranking problem into a multi-classification task, leveraging ordinal attributes to improve classification boundaries. Furthermore, in terms of fashion feature representation, we propose modeling not just individual items but also their combined effect as an outfit, providing a more holistic and nuanced fashion representation. We introduce the Fashionability3k dataset, comprising 3k image pairs (2398 ordered and 601 similar pairs) with objective relative fashion labels. Experiments on three datasets—our Fashionability3k and two public datasets—show that our method outperforms the baseline by nearly 1 % in ranking accuracy. In the user study, it achieved a 0.72 consistency with human subjective perception. Moreover, combining local and global visual features leads to additional performance gains, with an average improvement of 2.78 % in ordered pairs and 1.09 % in similar pairs. This is the first study to treat fashionability as an objective attribute for comparative analysis, validated through extensive experiments.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"62 5","pages":"Article 104205"},"PeriodicalIF":7.4,"publicationDate":"2025-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143901914","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

OVED-Rank: A ranking scheme to evaluate complex network spreaders’ influence through the concept of effective distance and orbital velocity oded - rank：一种通过有效距离和轨道速度的概念来评价复杂网络传播者影响力的排序方案

IF 7.4 1区管理学

Information Processing & Management Pub Date : 2025-05-02 DOI: 10.1016/j.ipm.2025.104201

Aman Ullah , Yahui Meng , J.F.F. Mendes

{"title":"OVED-Rank: A ranking scheme to evaluate complex network spreaders’ influence through the concept of effective distance and orbital velocity","authors":"Aman Ullah , Yahui Meng , J.F.F. Mendes","doi":"10.1016/j.ipm.2025.104201","DOIUrl":"10.1016/j.ipm.2025.104201","url":null,"abstract":"<div><div>This paper explores the influence of complex network spreaders, which is the most studied problems in network science. However, developing an efficient technique to handle this task remains challenging due to its NP-hard nature. Several traditional approaches to identifying the influence of complex network spreaders usually rely on algorithms for determining network paths that are more complex, such as Dijkstra’s algorithm or Bellman-Ford’s algorithm, which require significant computational resources and do not always take into account the location of nodes in a network. To cope with these issues, this paper proposes a new method called OVED-Rank, which is inspired by the orbital velocity formula for the influence of key spreaders in complex networks. It incorporates an advanced metric effective distance replacing traditional measures like Dijkstra’s distance to streamline computations and decrease processing times. OVED-Rank combines the degree of a node, the k-shell index, the number of triangles that form part of a node, and the length of paths connecting the nodes. Unlike traditional methods, OVED-Rank does not rely on the usual complex shortest-path algorithms. Instead, it uses effective distance, which makes the calculations easier and less complex. In addition, it improves predictability by taking into account the characteristics of neighboring nodes. The robustness and effectiveness of OVED-Rank are thoroughly vetted through rigorous testing on various network, including both synthetic setups and real-world undirected, unweighted networks. The experimental results are compelling, indicating that OVED-Rank not only meets but often exceeds the performance of existing methodologies.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"62 5","pages":"Article 104201"},"PeriodicalIF":7.4,"publicationDate":"2025-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143895253","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A Noise-Resistant Model for Graph-based Fraud Detection 基于图的欺诈检测抗噪声模型

IF 7.4 1区管理学

Information Processing & Management Pub Date : 2025-05-02 DOI: 10.1016/j.ipm.2025.104198

Zhengyang Liu, Hang Yu, Xiangfeng Luo

{"title":"A Noise-Resistant Model for Graph-based Fraud Detection","authors":"Zhengyang Liu, Hang Yu, Xiangfeng Luo","doi":"10.1016/j.ipm.2025.104198","DOIUrl":"10.1016/j.ipm.2025.104198","url":null,"abstract":"<div><div>Graph-based fraud detection is a critical task that identifies anomalous nodes that deviate from the majority of normal nodes within a graph. It can be applied in various practical situations, including but not limited to fake review detection, fraud transaction detection, and bot account detection. Current graph fraud detection models leverage popular Graph Neural Networks (GNNs) as their foundation, achieving significant success from the view of homogeneous and heterogeneous edges. However, these methods assume a sufficient proportion of completely accurate labeled nodes, overlooking the issue of noisy labels present in real-world scenarios. This can lead to significant performance degradation of current graph fraud detection methods. To address this challenge, we propose a Noise-Resistant Model for Graph Fraud Detection. First, we design a foundational graph fraud detection model from a spectral perspective to capture both homogeneous and heterogeneous information of nodes. Based on a conditional variational autoencoder(CVAE), we are able to obtain node features augmented from different perspectives. Next, nodes with noisy labels are trained alongside nodes with clean labels. Utilizing a self-supervised approach, noisy nodes with high prediction confidence that align with their labels are gradually incorporated to the training set. For nodes with lower confidence, we aim to learn better representations and gradually include more of them into the training set. With the augmented features generated by the CVAE, combined with a support set constructed from clean labels, we compute the consistency loss with adversarial strategies to ensure that features augmented from both normal and anomalous perspectives are brought closer to the relevant categories within the support set. Extensive experiments comparing our method with twelve state-of-the-art baselines on six real-world datasets – Amazon, Yelp, Elliptic, FDCompCN, T-Finance, and T-Social – showcase the superiority of our model.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"62 5","pages":"Article 104198"},"PeriodicalIF":7.4,"publicationDate":"2025-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143895252","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Enhancing the communication reliability for generative image steganography with diffusion model 利用扩散模型提高生成图像隐写的通信可靠性

IF 7.4 1区管理学

Information Processing & Management Pub Date : 2025-04-29 DOI: 10.1016/j.ipm.2025.104203

Mingliang Zhang, Xiangyang Luo, Pei Zhang, Yanmei Liu, Yi Zhang

{"title":"Enhancing the communication reliability for generative image steganography with diffusion model","authors":"Mingliang Zhang, Xiangyang Luo, Pei Zhang, Yanmei Liu, Yi Zhang","doi":"10.1016/j.ipm.2025.104203","DOIUrl":"10.1016/j.ipm.2025.104203","url":null,"abstract":"<div><div>Steganography, which conceals information within ordinary covers, can be applied in journalism, intelligence, and healthcare. It effectively avoids the risks of censorship, interception, and leakage, meeting the need for secure information transmission. Images generated by diffusion models are highly valued for their content authenticity and wide applicability, providing high-quality cover images for covert communication. However, existing methods face difficulties in generating stego images that conform to mainstream formats while ensuring the complete extraction of information. This may attract the attention of third parties and increase the risk of covert communication being detected and exposed. Therefore, we propose a diffusion model-based steganographic method to enhance communication reliability. The method first constructs a function that maps secret data to the latent vector space using non-continuous sub-intervals divided by the inverse cumulative distribution function and then generates an initial stego image using the diffusion model. Subsequently, an effective error detection mechanism is designed to address the potential loss of secret data during the quantization process based on the mapping rules and characteristics of the non-continuous sub-intervals. Finally, driven by the secret data, an adaptive quantization strategy is employed to iteratively correct the lost data based on the initial stego image and error detection information. Experimental results demonstrate that the proposed method can generate stego images in mainstream formats while demonstrating a consistently high extraction accuracy rate. Compared with methods that can ensure complete secret data extraction and use mainstream stego image formats, our method achieves state-of-the-art performance in embedding capacity.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"62 5","pages":"Article 104203"},"PeriodicalIF":7.4,"publicationDate":"2025-04-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143886882","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

MRLCD-A: Lag-aware alignment for multivariate time series forecasting in multiple scenarios MRLCD-A：多场景下多变量时间序列预测的滞后感知对齐

IF 7.4 1区管理学

Information Processing & Management Pub Date : 2025-04-27 DOI: 10.1016/j.ipm.2025.104191

Dezhi Sun , Jiwei Qin , Zihao Zhang , Xizhong Qin , Huiguo Zhang

引用次数: 0