EXPLAINER: Entity Resolution Explanations

Ebaid, Amr; Thirumuruganathan, Saravanan; Aref, Walid G.; Elmagarmid, Ahmed K.; Ouzzani, Mourad

doi:10.1109/icde.2019.00224

Cited by 22 publications

(14 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to better understand the matching decisions of different models, in the following we investigate the importance of words belonging to domain-specific word classes for the matching decisions. We base our analysis on the Mojito [10] framework, which adapts the LIME algorithm [28] for the use case of pairwise entity matching and is part of recent efforts to explain deep entity matching results [12,32]. LIME creates an explanation for a single matching decision as follows: The instance (pair of entity descriptions) is perturbed using word dropping and labels for all perturbed instances are queried from the model to be explained.…”

Section: Explaining Matching Decisionsmentioning

confidence: 99%

Dual-objective fine-tuning of BERT for entity matching

Peeters

Bizer

2021

Proc. VLDB Endow.

View full text Add to dashboard Cite

An increasing number of data providers have adopted shared numbering schemes such as GTIN, ISBN, DUNS, or ORCID numbers for identifying entities in the respective domain. This means for data integration that shared identifiers are often available for a subset of the entity descriptions to be integrated while such identifiers are not available for others. The challenge in these settings is to learn a matcher for entity descriptions without identifiers using the entity descriptions containing identifiers as training data. The task can be approached by learning a binary classifier which distinguishes pairs of entity descriptions for the same real-world entity from descriptions of different entities. The task can also be modeled as a multi-class classification problem by learning classifiers for identifying descriptions of individual entities. We present a dual-objective training method for BERT, called JointBERT, which combines binary matching and multi-class classification, forcing the model to predict the entity identifier for each entity description in a training pair in addition to the match/non-match decision. Our evaluation across five entity matching benchmark datasets shows that dual-objective training can increase the matching performance for seen products by 1% to 5% F1 compared to single-objective Transformer-based methods, given that enough training data is available for both objectives. In order to gain a deeper understanding of the strengths and weaknesses of the proposed method, we compare JointBERT to several other BERT-based matching methods as well as baseline systems along a set of specific matching challenges. This evaluation shows that JointBERT, given enough training data for both objectives, outperforms the other methods on tasks involving seen products, while it underperforms for unseen products. Using a combination of LIME explanations and domain-specific word classes, we analyze the matching decisions of the different deep learning models and conclude that BERT-based models are better at focusing on relevant word classes compared to RNN-based models.

show abstract

Section: Explaining Matching Decisionsmentioning

confidence: 99%

Dual-objective fine-tuning of BERT for entity matching

Peeters

Bizer

2021

Proc. VLDB Endow.

View full text Add to dashboard Cite

show abstract

“…Entity resolution tasks (ER or record linking) in NLP have seen much success with various post-hoc explanation techniques, especially with declarative induction methods. ExplainER by Ebaid et al uses Bayesian Rule List (BRL) to output declarative rules explaining the entity resolutions for global explanations [Eba+19].…”

Section: Declarative Inductionmentioning

confidence: 99%

Explainability in Time Series Forecasting, Natural Language Processing, and Computer Vision

Kamath

Liu

2021

Explainable Artificial Intelligence: An Introduction to Interpretable Machine Learning

View full text Add to dashboard Cite

Various domains such as computer vision, natural language processing, and time series analysis have extensively applied machine learning algorithms in recent years. This chapter will discuss the research and applications of the interpretable and explainable algorithms in this domain. We will start with a time series algorithm survey, starting from traditional interpretable statistical models to modern deep learning algorithms. Next, we discuss NLP applications and the role of interpretability. Finally, we cover computer vision and how explainability has been a focus of considerable research. We will present a case study in each domain where the reader can get practical and real-world insights. Time Series ForecastingForecasting, based on historical data, is one of the most critical applications of time series. There is a particular class of much easier problems, such as predicting daily temperature based on the last few days, while specific issues such as predicting volatility in the foreign exchange rates may be trickier. Understanding the factors, how they impact the target, seasonality, trend, etc., contribute to forecasting model quality. Time series modeling and forecasting have many parallels to the general machine learning process of training and predicting out-of-sample. This section will discuss the similarity and highlight the differences for time series modeling, especially from the explainability standpoint.

show abstract

“…Ribeiro et al (2016), Ribeiro et al (2018, Letham et al (2015) and Choudhary et al (2018) have proposed explainable systems for ER using local and if-then-else based global explanations. Ebaid et al (2019) is a tool that provides explanations at different granularity levels.…”

Section: Introductionmentioning

confidence: 99%

xER: An Explainable Model for Entity Resolution using an Efficient Solution for the Clique Partitioning Problem

Vadrevu¹,

Nagi²,

Xiong³

et al. 2021

Proceedings of the First Workshop on Trustworthy Natural Language Processing

View full text Add to dashboard Cite

In this paper, we propose a global, selfexplainable solution to solve a prominent NLP problem: Entity Resolution (ER). We formulate ER as a graph partitioning problem. Every mention of a real-world entity is represented by a node in the graph, and the pairwise similarity scores between the mentions are used to associate these nodes to exactly one clique, which represents a real-world entity in the ER domain. In this paper, we use Clique Partitioning Problem (CPP), which is an Integer Program (IP) to formulate ER as a graph partitioning problem and then highlight the explainable nature of this method. Since CPP is NP-Hard, we introduce an efficient solution procedure, the xER algorithm, to solve CPP as a combination of finding maximal cliques in the graph and then performing generalized set packing using a novel formulation. We discuss the advantages of using xER over the traditional methods and provide the computational experiments and results of applying this method to ER data sets.

show abstract

EXPLAINER: Entity Resolution Explanations

Cited by 22 publications

References 9 publications

Dual-objective fine-tuning of BERT for entity matching

Dual-objective fine-tuning of BERT for entity matching

Explainability in Time Series Forecasting, Natural Language Processing, and Computer Vision

xER: An Explainable Model for Entity Resolution using an Efficient Solution for the Clique Partitioning Problem

Contact Info

Product

Resources

About