Bootstrapping Multilingual Relation Discovery Using English Wikipedia and Wikimedia-Induced Entity Extraction

Spatial relation extraction is the process of identifying geographic entities from text and determining their corresponding spatial relations. Traditional spatial relation extraction mainly uses rule-based pattern matching, supervised learning-based or unsupervised learning-based methods. However, these methods suffer from poor time-sensitive, high labor cost and high dependence on large-scale data. With the development of pre-trained language models greatly alleviating the shortcomings of traditional methods, supervised learning methods incorporating pre-trained language models have become the mainstream relation extraction methods. Pipeline extraction and joint extraction, as the two most dominant ideas of relation extraction, both have obtained good performance on different datasets, and whether to share the contextual information of entities and relations is the main differences between the two ideas. In this paper, we compare the performance of two ideas oriented to spatial relation extraction based on Chinese corpus data in the field of geography and verify which method based on pre-trained language models is more suitable for Chinese spatial relation extraction. We fine-tuned the hyperparameters of the two models to optimize the extraction accuracy before the comparison experiments. The results of the comparison experiments show that pipeline extraction performs better than joint extraction of spatial relation extraction for Chinese text data with sentence granularity, because different tasks have different focus on contextual information, and it is difficult to take account into the needs of both tasks by sharing contextual information. In addition, we further compare the performance of the two models with the rule-based template approach in extracting topological, directional and distance relations, summarize the shortcomings of this experiment and provide an outlook for future work.

Distantly supervised relation extraction has been widely applied in knowledge base construction due to its less requirement of human efforts. However, the automatically established training datasets in distant supervision contain low-quality instances with noisy words and overlapped relations, introducing great challenges to the accurate extraction of relations. To address this problem, we propose a novel Regularized Attentive Capsule Network (RA-CapNet) to better identify highly overlapped relations in each informal sentence. To discover multiple relation features in an instance, we embed multi-head attention into the capsule network as the low-level capsules, where the subtraction of two entities acts as a new form of relation query to select salient features regardless of their positions. To further discriminate overlapped relation features, we devise disagreement regularization to explicitly encourage the diversity among both multiple attention heads and low-level capsules. Extensive experiments conducted on widely used datasets show that our model achieves significant improvements in relation extraction.

Distantly supervised relation extraction has been widely applied in knowledge base construction due to its less requirement of human efforts. However, the automatically established training datasets in distant supervision contain low-quality instances with noisy words and overlapped relations, introducing great challenges to the accurate extraction of relations. To address this problem, we propose a novel Regularized Attentive Capsule Network (RA-CapNet) to better identify highly overlapped relations in each informal sentence. To discover multiple relation features in an instance, we embed multi-head attention into the capsule network as the low-level capsules, where the subtraction of two entities acts as a new form of relation query to select salient features regardless of their positions. To further discriminate overlapped relation features, we devise disagreement regularization to explicitly encourage the diversity among both multiple attention heads and low-level capsules. Extensive experiments conducted on widely used datasets show that our model achieves significant improvements in relation extraction.

Natural Language Processing techniques play an essential role in extraction of necessary information for ontology construction from unstructured text. Identifying syntactic constituents and their dependencies in a sentence, boost the information extraction from natural language text. Main ingredients required for ontology construction are required to be extracted from text in the form of entities and relations between them. Language dependency constructs that express the binary dependencies of the lexical terms in a sentence can be considered as good indicators in identifying binary relationships existing between entities. In this paper we describe that how the typed dependencies produced by a natural language parser are used by a multi agent system to generate rules for relation extraction between two identified entities. The typed dependencies produced by parsing are processed to eliminate unnecessary dependencies and make them more appropriate to be used in rule learning for relation extraction. All the relations derived are expressed as predicate expressions of two entities. We evaluate our agent system by applying it on number of wikipedia web pages from the domain of birds.

Although Large language Model (LLM)-powered information extraction (IE) systems have shown impressive capabilities, current fine-tuning paradigms face two major limitations: high training costs and difficulties in aligning with LLM preferences. To address these issues, we propose a novel universal IE paradigm—the Self-Correcting Iterative Refinement (SCIR) framework—along with a Multi-task Bilingual (Chinese-English) Self-Correcting (MBSC) dataset containing over 100,000 entries. The SCIR framework achieves plug-and-play compatibility with existing LLMs and IE systems through its Dual-Path Self-Correcting module and feedback-driven optimization, thereby significantly reducing training costs. Concurrently, the MBSC dataset tackles the challenge of preference alignment by indirectly distilling GPT-4's capabilities into IE result detection models. Experimental results demonstrate that SCIR outperforms state-of-the-art IE methods across three key tasks— named entity recognition, relation extraction, and event extraction—achieving a 5.27 percent average improvement in span-based Micro-F1 while reducing training costs by 87 percent compared to baseline approaches. These advancements not only enhance the flexibility and accuracy of IE systems but also pave the way for lightweight and efficient IE paradigms.

In the contemporary digital era, the exponential increase in unstructured and semi-structured data has made information extraction a cornerstone of modern data-driven research and application. The ability to transform such raw information into structured knowledge is crucial for enabling later downstream tasks. While traditional rule-based and statistical approaches to information extraction have demonstrated success in narrow, well-defined tasks, they lack the scalability and adaptability required to address the vastness and variability of present-day data. Conversely, deep neural models and especially large language models have shown remarkable capabilities in language understanding, yet they remain constrained by high computational costs and susceptibility to hallucination. <br> This thesis explores the unification of various symbolic, statistical, and neural paradigms into a cohesive hybrid framework. The central hypothesis is that by combining the strengths of data-driven representation learning with structural, rule-based, and multimodal knowledge, one can achieve information extraction systems that are more accurate, efficient, and reliable than their monolithic counterparts. To test this hypothesis, the thesis investigates a range of hybrid architectures across five key application domains. <br> In the financial domain, a hybrid contradiction detection framework integrates syntactic pre-training with transformer-based representations and clustering algorithms to identify inconsistencies within large-scale financial reports. For named entity recognition, the <em>iNERD</em> algorithm introduces rule-based constraints to guide large language models, producing syntactically valid, hallucination-free entity extractions. Thereafter, the anonymisation study leverages knowledge distillation to compress the language understanding capabilities of large decoder-only models into lightweight encoder-only architectures, enabling secure and efficient text anonymisation. In relation extraction, this work presents <em>KPI-BERT</em> and the open-source <em>KPI-EDGAR</em> dataset, combining contextual embedding models with recurrent layers and noise-based regularisation to extract key performance indicators from financial documents. Extending beyond text, the final empirical contribution introduces a multimodal dementia detection framework that fuses linguistic and acoustic representations, offering a robust approach to early, non-invasive diagnosis. <br> Together, these studies provide compelling evidence that hybrid representation learning constitutes an important paradigm for modern information extraction. This research demonstrates that hybrid systems can achieve higher precision, stronger generalisability, and improved efficiency while remaining adaptable to real-world constraints. The findings of this thesis therefore advance the field towards more trustworthy, sustainable, and application-ready artificial intelligence.

The schematization of knowledge, including the extraction of entities and relations from documents, poses significant challenges to traditional approaches because of the document's ambiguity, heterogeneity, and high cost domainspecific training.Although Large Language Models (LLMs) allow for extraction without prior training on the dataset, the requirement of fine-tuning along with low precision, especially in relation extraction, serves as an obstacle.In absence of domain-specific training, we present a new zero-shot ensemble approach using DeepSeek-R1-Distill-Llama-70B, Llama-3.3-70B, and Qwen-2.5-32B.Our key innovation is a two-stage pipeline that first consolidates high-confidence entities through ensemble techniques, then leverages Qwen-2.5-32B with engineered prompts to generate precise semantic triples.This approach effectively resolves the low precision problem typically encountered in relation extraction.Experiments demonstrate significant gains in both accuracy and efficiency across diverse domains, with our method ranking in the top 2 on the official leaderboard in Shared Task-IV of The 1st Joint Workshop on Large Language Models and Structure Modeling.This competitive performance validates our approach as a compelling solution for practitioners seeking robust document-level information extraction without the burden of task-specific fine-tuning.

Multi-tiered storage, where each tier comprises one type of storage device, e.g., SSD, HDD, is a commonly used approach to achieve both high performance and cost efficiency in large-scale systems that need to store data with vastly different access characteristics. By aligning the access characteristics of the data to the characteristics of the storage devices, higher performance can be achieved for any given cost. This article presents ExaPlan, a method to determine both the data-to-tier assignment and the number of devices in each tier that minimize the system's mean response time for a given budget and workload. In contrast to other methods that constrain or minimize the system load, ExaPlan directly minimizes the system's mean response time estimated by a queueing model. Minimizing the mean response time is typically intractable as the resulting optimization problem is both non-convex and combinatorial in nature. ExaPlan circumvents this intractability by introducing a parameterized data-placement approach that makes it a highly scalable method that can be easily applied to exascale systems. Through experiments that use parameters from real-world storage systems, such as CERN and LOFAR, it is demonstrated that ExaPlan provides solutions that yield lower mean response times than previous works. It is also capable of determining a data-to-tier assignment both at the level of files and at the level of fixed-size extents. For some of the workloads evaluated, file-level placement exhibited a significant performance improvement over extent-level placement.

Meaningful Information extraction is an extremely important and challenging task due to the ever growing size of data. Training and evaluating automated systems for the task requires annotated datasets which are rarely available because of the great amount of human effort and time required for annotating data. The dataset described in this manuscript, CustFRE, is meant for systems that learn extracting family relations from text. Sentences having at least two persons have been collected from the internet. The texts are first processed using Stanford's NLP pipeline for basic NLP tagging. Next, a team of natural language processing experts annotated the dataset. All family relations among persons in the texts have been annotated, or a no_relation is annotated if no family relation between two persons can be inferred from the text. After annotation, the dataset was verified by an NLP expert for completeness and correctness. CustFRE contains in total 2,716 annotations. The dataset can be used by information extraction researchers as a benchmark for evaluating their systems, and can also be used for training and evaluating family relation extraction systems.

In the traditional English learning mode of vocational college, teachers and students spend a lot of time on the basic knowledge of English. Due to the singularity of teaching methods and the lack of effective student self-learning platform, many students consider that English learning is boring and the actual learning effect is poor[1-2]. As an emerging edge discipline, artificial intelligence covers a series of topics such as image recognition, pattern recognition, robotics, natural language understanding and expert systems. The simulation and application of human language intelligence is an essential progress of artificial intelligence technology[3]. This paper will analyze the connotation of artificial intelligence technology and its principles of natural language understanding, and sort out the application advantages of artificial intelligence in education field. Finally, it puts forward practical measures of adopting artificial intelligence tools in input and output stages of English language learning[4-5].

There are limited training options for audiovisual archivists, with most formal courses centred in Europe or the United States of America, but high costs can prevent people working in audiovisual archives from accessing these opportunities. However, there are significant collections of audiovisual heritage spread across the globe, not the least in Southeast Asia and the Pacific region, that are at risk of loss due to a number of factors, including sta competencies. In 1996 audiovisual archivists formed the Southeast Asia–Pacific Audiovisual Archive Association (SEAPAVAA) to advocate on their behalf and to provide networking and other assistance to develop and sustain their respective collections. A key part of SEAPAVAA’s work has been to provide training. Over the past 20 years the association has developed and delivered educational programmes on all aspects of audiovisual archiving. Over this time its trainers have developed an analytical approach to prioritizing needs and optimizing delivery methods in a region that has many distinct languages and cultures and where one size does not fit all. This paper looks at how SEAPAVAA went about discovering those needs and developing training priorities around them.

Scientists depend on complex computational systems that are often ineliminably opaque, to the detriment of our ability to give scientific explanations and detect artifacts. Some philosophers have suggested treating opaque systems instrumentally, but computer scientists developing strategies for increasing transparency are correct in finding this unsatisfying. Instead, I propose an analysis of transparency as having three forms: transparency of the algorithm, the realization of the algorithm in code, and the way that code is run on particular hardware and data. This targets the transparency most useful for a task, avoiding instrumentalism by providing partial transparency when full transparency is impossible.

To alleviate human efforts from obtaining large-scale annotations, Semi-Supervised Relation Extraction methods aim to leverage unlabeled data in addition to learning from limited samples. Existing self-training methods suffer from the gradual drift problem, where noisy pseudo labels on unlabeled data are incorporated during training. To alleviate the noise in pseudo labels, we propose a method called MetaSRE, where a Relation Label Generation Network generates quality assessment on pseudo labels by (meta) learning from the successful and failed attempts on Relation Classification Network as an additional meta-objective. To reduce the influence of noisy pseudo labels, MetaSRE adopts a pseudo label selection and exploitation scheme which assesses pseudo label quality on unlabeled samples and only exploits high-quality pseudo labels in a self-training fashion to incrementally augment labeled samples for both robustness and accuracy. Experimental results on two public datasets demonstrate the effectiveness of the proposed approach.

Most information extraction systems either use hand written extraction patterns or use a machine learning algorithm that is trained on a manually annotated corpus. Both of these approaches require massive human effort and hence prevent information extraction from becoming more widely applicable. In this paper we present URES (Unsupervised Relation Extraction System), which extracts relations from the Web in a totally unsupervised way. It takes as input the descriptions of the target relations, which include the names of the predicates, the types of their attributes, and several seed instances of the relations. Then the system downloads from the Web a large collection of pages that are likely to contain instances of the target relations. From those pages, utilizing the known seed instances, the system learns the relation patterns, which are then used for extraction. We present several experiments in which we learn patterns and extract instances of a set of several common IE relations, comparing several pattern learning and filtering setups. We demonstrate that using simple noun phrase tagger is sufficient as a base for accurate patterns. However, having a named entity recognizer, which is able to recognize the types of the relation attributes significantly, enhances the extraction performance. We also compare our approach with KnowItAll's fixed generic patterns.

Dynamic logic-based attack graph for risk assessment in complex computer systems