Entity Resolution (ER) is a fundamental task in data integration, particularly in streaming environments where entities arrive continuously and decisions must be made under strict time constraints. Existing approaches primarily optimize efficiency and accuracy, but often overlook the interpretability of matching decisions and may propagate group-level disparities. To address these limitations, this paper introduces X-TREATS, a streaming-oriented ER workflow that integrates pair-level explanations directly into the resolution pipeline, jointly combining similarity, fairness, and explanation signals during ranking step. The proposed approach is evaluated on real-world datasets under incremental processing, assessing effectiveness, fairness, and explainability. Experimental results show that X-TREATS substantially improves interpretability, increasing the Explanation Score by up to 20%, while reducing group-level disparities by 40–70% and preserving high matching precision across all datasets. These findings demonstrate the practical benefits of integrating explanation-aware mechanisms into real-time ER pipelines.

Currently, the growing proliferation of information systems generates large volumes of data continuously, stemming from a variety of sources such as web platforms, social networks, and multiple devices. These data, often lacking a defined schema, require an initial process of consolidation and cleansing before analysis and knowledge extraction can occur. In this context, Entity Resolution (ER) plays a crucial role, facilitating the integration of knowledge bases and identifying similarities among entities from different sources. However, the traditional ER process is computationally expensive, and becomes more complicated in the streaming context where the data arrive continuously. Moreover, there is a lack of studies involving fairness and ER, which is related to the absence of discrimination or bias. In this sense, fairness criteria aim to mitigate the implications of data bias in ER systems, which requires more than just optimizing accuracy, as traditionally done. Considering this context, this work presents TREATS, a schema-agnostic and fairness-aware ER workflow developed for managing streaming data incrementally. The proposed fairness-aware ER framework tackles constraints across various groups of interest, presenting a resilient and equitable solution to the related challenges. Through experimental evaluation, the proposed techniques and heuristics are compared against state-of-the-art approaches over five real-world data source pairs, in which the results demonstrated significant improvements in terms of fairness, without degradation of effectiveness and efficiency measures in the streaming environment. In summary, our contributions aim to propel the ER field forward by providing a workflow that addresses both technical challenges and ethical concerns.

Property Graphs are a versatile and expressive data model that has gained widespread adoption due to their flexibility in supporting labeled and attributed nodes and edges. They are well-established in research communities and are becoming widespread in companies and organizations across various sectors. They have been boosted by a fervent ISO/IEC standardization activity, leading to dedicated query and schema languages. While the current standards are still evolving, opportunities remain to enrich them with features such as composability. The plethora of existing query languages reflects a rich and diverse ecosystem, which ongoing unification efforts aim to align. This tutorial aims to deepen the understanding of Property Graph standards by showcasing their strengths, highlighting recent unification efforts, clarifying the central role of schema constraints, and exploring the rich landscape of research and industrial opportunities shaping the future of graph data management.

Data practitioners often need to sample their datasets to produce representative subsets for their downstream tasks. Unfortunately, real-world datasets frequently contain duplicates, whose presence biases sampling and impacts the quality of the produced subsets, hence the outcome of downstream tasks. While deduplication is therefore fundamental, performing it on the entire dataset to run sampling on its cleaned version might be prohibitively expensive in terms of time and resources. Thus, we recently introduced RadlER, a solution to perform deduplicated sampling on-demand , i.e., to produce a clean sample of a dirty dataset incrementally, according to a target distribution of some subpopulations, by focusing the cleaning effort only on entities required to appear in the sample. In this demonstration, we interactively show how RadlER can support practitioners in their data science pipelines, allowing them to save a relevant amount of time and resources.

Data practitioners often sample their datasets to produce representative subsets for their downstream tasks. When entities in a dataset can be partitioned into multiple groups, stratified sampling is commonly used to produce subsets that match a target group distribution, e.g., to select a balanced subset for training a machine learning model. However, real-world data frequently contains duplicates — multiple representations of the same real-world entity — that can bias sampling, necessitating deduplication. We define deduplicated sampling as the task of producing a clean sample of a dirty dataset according to a target group distribution. The naïve approach to deduplicated sampling would first deduplicate the entire dataset upfront, then perform sampling ex post. However, that approach might be prohibitively expensive for large datasets and time/resource constraints. Deduplicated sampling ondemand with RadlER is a novel approach to produce a clean sample by focusing the cleaning effort only on entities required to appear in that sample. Our experimental evaluation, performed on multiple datasets from different domains, demonstrates that RadlER consistently outperforms baseline approaches, providing data scientists with an efficient solution to quickly produce a clean sample of a dirty dataset according to a target group distribution.

The pervasive spread of disinformation across social media platforms has become a significant global challenge, disrupting democratic processes, undermining public trust, and fueling societal polarization. Existing approaches often neglect the dynamic and structural mechanisms that drive the spread and adoption of false narratives. This article leverages the well-established principles and methodologies of Statistical Mechanics and introduces a dynamic Mean-Field framework to model the evolution of disinformation within social networks. The framework introduces innovative elements, including heterogeneous coupling strengths to capture diverse social influences among network users, memory effects to account for cognitive inertia or belief re-evaluation and a three-state Potts model to represent polarization and neutrality in opinion dynamics. It employs the concept of effective fields to integrate external disinformation campaigns, facilitating a detailed analysis of critical thresholds and phase transitions. Monte Carlo simulations are performed to further illustrate the transient and equilibrium dynamics of belief adoption and rejection. Our findings provide actionable insights for the disinformation spread and offer a theoretical foundation for designing targeted interventions to mitigate its harmful effects on societies.

The rise of financial crime that has been observed in recent years has created an increasing concern around the topic and many people, organizations and governments are more and more frequently trying to combat it. Despite the increase of interest in this area, there is a lack of specialized datasets that can be used to train and evaluate works that try to tackle those problems. This article proposes a new micro-benchmark dataset for algorithms and models that identify individuals and organizations, and their multiple writings, in news articles, and presents an approach that assists in its creation. Experimental efforts are also reported, using this dataset, to identify individuals and organizations in financial-crime-related articles using various low-billion parameter Large Language Models (LLMs). For these experiments, standard metrics (Accuracy, Precision, Recall, F1 Score) are reported and various prompt variants comprising the best practices of prompt engineering are tested. In addition, to address the problem of ambiguous entity mentions, a simple, yet effective LLM-based disambiguation method is proposed, ensuring that the evaluation aligns with reality. Finally, the proposed approach is compared against a widely used stateof-the-art open-source baseline, showing the superiority of the proposed method.

Infusing autonomous artificial systems with knowledge about the physical world they inhabit is a critical and long-held aim for the Artificial Intelligence community. Training systems with relevant data is a typical approach; however, finding the data required is not always possible, especially when much of this knowledge is commonsense. In this paper, we present a comparison of topology-based and semantics-based methods for extracting information about object-action and object-state association relations from knowledge graphs, such as ConceptNet, WordNet, ATOMIC, YAGO, WebChild and DBpedia. Moreover, we propose a novel method for extracting information about object-action and object-state associations from knowledge graphs. Our method is composed of a set of techniques for locating, enriching, evaluating, cleaning and exposing knowledge from such resources, relying on semantic similarity methods. Some important aspects of our method are the flexibility in deciding how to deal with the noise that exists in the data, and the capability to determine the importance of a path through training, rather than through manual annotation.

Entity Resolution has been an active research topic for the last three decades, with numerous algorithms proposed in the literature. However, putting them into practice is often a complex task that requires implementing, combining and configuring complementary individual algorithms into comprehensive end-to-end workflows. To facilitate this process, we are developing pyJedAI, a novel system that provides a unifying framework for any type of main works in the field (i.e., both unsupervised and learning-based ones). Our vision is to facilitate both novice and expert users to use and combine these algorithms through a series of principled approaches for automatically configuring and benchmarking end-to-end pipelines.

Abstract In recent years, we have witnessed the proliferation of knowledge graphs (KG) in various domains, aiming to support applications like question answering, recommendations, etc. A frequent task when integrating knowledge from different KGs is to find which subgraphs refer to the same real-world entity, a task largely known as the Entity Alignment. Recently, embedding methods have been used for entity alignment tasks, that learn a vector-space representation of entities which preserves their similarity in the original KGs. A wide variety of supervised, unsupervised, and semi-supervised methods have been proposed that exploit both factual (attribute based) and structural information (relation based) of entities in the KGs. Still, a quantitative assessment of their strengths and weaknesses in real-world KGs according to different performance metrics and KG characteristics is missing from the literature. In this work, we conduct the first meta-level analysis of popular embedding methods for entity alignment, based on a statistically sound methodology. Our analysis reveals statistically significant correlations of different embedding methods with various meta-features extracted by KGs and rank them in a statistically significant way according to their effectiveness across all real-world KGs of our testbed. Finally, we study interesting trade-offs in terms of methods’ effectiveness and efficiency.

Abstract Entity resolution (ER) is the task of finding records that refer to the same real-world entities. A common scenario, which we refer to as Clean-Clean ER, is to resolve records across two clean sources (i.e., they are duplicate-free and contain one record per entity). Matching algorithms for Clean-Clean ER yield bipartite graphs, which are further processed by clustering algorithms to produce the end result. In this paper, we perform an extensive empirical evaluation of eight bipartite graph matching algorithms that take as input a bipartite similarity graph and provide as output a set of matched records. We consider a wide range of matching algorithms, including algorithms that have not previously been applied to ER, or have been evaluated only in other ER settings. We assess the relative performance of these algorithms with respect to accuracy and time efficiency over ten established real-world data sets, from which we generated over 700 different similarity graphs. Our results provide insights into the relative performance of these algorithms and guidelines for choosing the best one, depending on the data at hand.

There is an urgent call to detect and prevent "biased data" at the earliest possible stage of the data pipelines used to build automated decision-making systems. In this paper, we are focusing on controlling the data bias in entity resolution (ER) tasks aiming to discover and unify records/descriptions from different data sources that refer to the same real-world entity. We formally define the ER problem with fairness constraints ensuring that all groups of entities have similar chances to be resolved. Then, we introduce FairER, a greedy algorithm for solving this problem for fairness criteria based on equal matching decisions. Our experiments show that FairER achieves similar or higher accuracy against two baseline methods over 7 datasets, while guaranteeing minimal bias.

Enterprises are creating domain-specific knowledge graphs by curating and integrating their business data from multiple sources. Ontologies provide a semantic abstraction for such knowledge graphs to describe their data in terms of the entities involved and their relationships. There has been a lot of effort to build systems that enable efficient querying over knowledge graphs, represented as property graphs. However the problem of schema optimization in the property graph setting has been largely ignored. In this work, we show that graph schema design has significant impact on query performance, and propose two algorithms to generate an optimized property graph schema from the domain ontology. To the best of our knowledge, we are the first to present an ontology-driven approach for property graph schema optimization. The rich semantic relationships in an ontology contain a variety of opportunities to reduce edge traversals and consequently improve the graph query performance. Our experimental study with two real-world knowledge graphs shows that our algorithms produce high-quality schemas, achieving up to 2 orders of magnitude speed-up compared to alternative schema designs.

Medical ontologies are widely used to describe and organize medical terminologies and to support many critical applications on healthcare databases. These ontologies are often manually curated (e.g., UMLS, SNOMED CT, and MeSH) by medical experts. Medical databases, on the other hand, are often created by database administrators, using different terminology and structures. The discrepancies between medical ontologies and databases compromise interoperability between them. Data to ontology matching is the process of finding semantic correspondences between tables in databases to standard ontologies. Existing solutions such as ontology matching have mostly focused on engineering features from terminological, structural, and semantic model information extracted from the ontologies. However, this is often labor intensive and the accuracy varies greatly across different ontologies. Worse yet, the ontology capturing a medical database is often not given in practice. In this paper, we propose MEDTO, a novel end-to-end framework that consists of three innovative techniques: (1) a lightweight yet effective method that bootstrap a semantically rich ontology from a given medical database, (2) a hyperbolic graph convolution layer that encodes hierarchical concepts in the hyperbolic space, and (3) a heterogeneous graph layer that encodes both local and global context information of a concept. Experiments on two real-world medical datasets matching against SNOMED CT show significant improvements compared to the state-of-the-art methods. MEDTO also consistently achieves competitive results on a benchmark from the Ontology Alignment Evaluation Initiative.

Medical knowledge bases (KBs), distilled from biomedical literature and regulatory actions, are expected to provide high-quality information to facilitate clinical decision making. Entity disambiguation (also referred to as entity linking) is considered as an essential task in unlocking the wealth of such medical KBs. However, existing medical entity disambiguation methods are not adequate due to word discrepancies between the entities in the KB and the text snippets in the source documents. Recently, graph neural networks (GNNs) have proven to be very effective and provide state-of-the-art results for many real-world applications with graph-structured data. In this paper, we introduce ED-GNN based on three representative GNNs (GraphSAGE, R-GCN, and MAGNN) for medical entity disambiguation. We develop two optimization techniques to fine-tune and improve ED-GNN. First, we introduce a novel strategy to represent entities that are mentioned in text snippets as a query graph. Second, we design an effective negative sampling strategy that identifies hard negative samples to improve the model's disambiguation capability. Compared to the best performing state-of-the-art solutions, our ED-GNN offers an average improvement of 7.3% in terms of F1 score on five real-world datasets.

Recent advances in natural language understanding and processing resulted in renewed interest in natural language based interfaces to data, which provide an easy mechanism for non-technical users to access and query the data. While early systems only allowed simple selection queries over a single table, some recent work supports complex BI queries, with many joins and aggregation, and even nested queries. There are various approaches in the literature for interpreting user's natural language query. Rule-based systems try to identify the entities in the query, and understand the intended relationships between those entities. Recent years have seen the emergence and popularity of neural network based approaches which try to interpret the query holistically, by learning the patterns. In this tutorial, we will review these natural language interface solutions in terms of their interpretation approach, as well as the complexity of the queries they can generate. We will also discuss open research challenges.

Domain-specific knowledge bases (KB), carefully curated from various data sources, provide an invaluable reference for professionals. Conversation systems make these KBs easily accessible to professionals and are gaining popularity due to recent advances in natural language understanding and AI. Despite the increasing use of various conversation systems in open-domain applications, the requirements of a domain-specific conversation system are quite different and challenging. In this paper, we propose an ontology-based conversation system for domain-specific KBs. In particular, we exploit the domain knowledge inherent in the domain ontology to identify user intents, and the corresponding entities to bootstrap the conversation space. We incorporate the feedback from domain experts to further refine these patterns, and use them to generate training samples for the conversation model, lifting the heavy burden from the conversation designers. We have incorporated our innovations into a conversation agent focused on healthcare as a feature of the IBM Micromedex product.

Entity Resolution (ER) aims to identify different descriptions in various Knowledge Bases (KBs) that refer to the same entity. ER is challenged by the Variety, Volume and Veracity of entity descriptions published in the Web of Data. To address them, we propose the MinoanER framework that simultaneously fulfills full automation, support of highly heterogeneous entities, and massive parallelization of the ER process. MinoanER leverages a token-based similarity of entities to define a new metric that derives the similarity of neighboring entities from the most important relations, as they are indicated only by statistics. A composite blocking method is employed to capture different sources of matching evidence from the content, neighbors, or names of entities. The search space of candidate pairs for comparison is compactly abstracted by a novel disjunctive blocking graph and processed by a non-iterative, massively parallel matching algorithm that consists of four generic, schema-agnostic matching rules that are quite robust with respect to their internal configuration. We demonstrate that the effectiveness of MinoanER is comparable to existing ER tools over real KBs exhibiting low Variety, but it outperforms them significantly when matching KBs with high Variety.

Entity Resolution (ER) aims to identify different descriptions in various Knowledge Bases (KBs) that refer to the same entity. ER is challenged by the Variety, Volume and Veracity of descriptions published in the Web of Data. To address them, we propose the MinoanER framework that fulfills full automation and support of highly heterogeneous entities. MinoanER leverages a token-based similarity of entities to define a new metric that derives the similarity of neighboring entities from the most important relations, indicated only by statistics. For high efficiency, similarities are computed from a set of schema-agnostic blocks and processed in a non-iterative way that involves four threshold-free heuristics. We demonstrate that the effectiveness of MinoanER is comparable to existing ER tools over real KBs exhibiting low heterogeneity in terms of entity types and content. Yet, MinoanER outperforms state-of-the-art ER tools when matching highly heterogeneous KBs.

Entity resolution aims to identify descriptions of the same entity within or across knowledge bases. In this work, we provide a comprehensive and cohesive overview of the key research results in the area of entity resolution. We are interested in frameworks addressing the new challenges in entity resolution posed by the Web of data in which real world entities are described by interlinked data rather than documents. Since such descriptions are usually partial, overlapping and sometimes evolving, entity resolution emerges as a central problem both to increase dataset linking, but also to search the Web of data for entities and their relations. We focus on Web-scale blocking, iterative and progressive solutions for entity resolution. Specifically, to reduce the required number of comparisons, blocking is performed to place similar descriptions into blocks and executes comparisons to identify matches only between descriptions within the same block. To minimize the number of missed matches, an iterative entity resolution process can exploit any intermediate results of blocking and matching, discovering new candidate description pairs for resolution. Finally, we overview works on progressive entity resolution, which attempt to discover as many matches as possible given limited computing budget, by estimating the matching likelihood of yet unresolved descriptions, based on the matches found so far.

Entity resolution aims to identify descriptions of the same entity within or across knowledge bases. In this work, we present the Minoan ER platform for resolving entities described by linked data in the Web (e.g., in RDF). To reduce the required number of comparisons, Minoan ER performs blocking to place similar descriptions into blocks and executes comparisons to identify matches only between descriptions within the same block. Moreover, it explores in a pay-as-you-go fashion any intermediate results of matching to obtain similarity evidence of entity neighbors and discover new candidate description pairs for resolution.

Entity resolution constitutes a crucial task for many applications, but has an inherently quadratic complexity. Typically, it scales to large volumes of data through blocking: similar entities are clustered into blocks so that it suffices to perform comparisons only within each block. Meta-blocking further increases efficiency by cleaning the overlapping blocks from unnecessary comparisons. However, even Meta-blocking can be time-consuming: applying it to blocks with 7.4 million entities and 2.21011 comparisons takes almost 8 days on a modern high-end server. In this paper, we parallelize Meta-blocking based on MapReduce. We propose a simple strategy that explicitly creates the core concept of Meta-blocking, the blocking graph. We then describe an advanced strategy that creates the blocking graph implicitly, reducing the overhead of data exchange. We also introduce a load balancing algorithm that distributes the computationally intensive workload evenly among the available compute nodes. Our experimental analysis verifies the superiority of our advanced strategy and demonstrates an almost linear speedup for all meta-blocking techniques with respect to the number of available nodes.

In the Web of data, entities are described by interlinked data rather than documents on the Web. In this work, we focus on entity resolution in the Web of data, i.e., identifying descriptions that refer to the same real-world entity. To reduce the required number of pairwise comparisons, methods for entity resolution perform blocking as a pre-processing step. A blocking technique places similar entity descriptions into blocks and executes comparisons only between descriptions within the same block. We experimentally evaluate blocking techniques proposed for the Web of data and present dataset characteristics that determine the effectiveness and efficiency of such methods. Furthermore, we analyze the characteristics of the missed matching entity descriptions and examine different types of links that blocking techniques can potentially identify.

Top-k is a well-studied problem in the literature, due to its wide spectrum of applications, like information retrieval, database querying, Web search and data mining. In the big data era, the volume of the data and their velocity, call for efficient parallel solutions that overcome the restricted resources of a single machine. Our motivating application is recommenders, which typically deal with big numbers of users and items, but other applications might benefit as well, like keyword search. In this paper, we propose a parallel top-k MapReduce algorithm that, unlike existing MapReduce solutions, manages to handle cases in which the k results do not fit in memory.