1 Introduction

A recent survey by IBMFootnote 1 suggests that more than 2.5 quintillion bytes of data are produced on the Web every day. Entity Linking (EL), also known as Named Entity Disambiguation (NED), is one of the most important Natural Language Processing (NLP) techniques for extracting knowledge automatically from this huge amount of data. The goal of an EL approach is as follows: Given a piece of text, a reference knowledge base K and a set of entity mentions in that text, map each entity mention to the corresponding resource in K [4]. A large number of challenges has to be addressed while performing a disambiguation. For instance, a given resource can be referred to using different labels due to phenomena such as synonymy, acronyms or typos. For example, New York City, NY and Big Apple are all labels for the same entity. Also, multiple entities can share the same name due to homonymy and ambiguity. For example, both the state and the city of Rio de Janeiro are called Rio de Janeiro.

Despite the complexity of the task, EL approaches have recently achieved increasingly better results by relying on trained machine learning models [6]. A portion of these approaches claim to be multilingual and most of them rely on models which are trained on English corpora with cross-lingual dictionaries. However, MAG (Multilingual AGDISTIS) [4] showed that the underlying models being trained on English corpora make them prone to failure when migrated to a different language. Additionally, these approaches hardly make their models or data available on more than three languages [6]. The new version of MAG (which is the quintessence of this demo) provides support for 40 different languages using sophisticated indicesFootnote 2. For the sake of server space, we deployed MAG-based web services for 9 languages and offer the other 31 languages for download. Additionally, we provide an English index using Wikidata to show the knowledge-base agnosticism of MAG. During the demo, we will show how to use the web services as well as MAG’s user interface.

2 MAG Entity Linking System

MAG’s EL process comprises two phases, namely an offline and an online phase. The sub-indices (which are generated during the offline phase) consist of surface forms, person names, rare references, acronyms and context information. During the online phase, the EL is carried out in two steps: (1) candidate generation and (2) disambiguation. The goal of the candidate generation step is to retrieve a tractable number of candidates for each mention. These candidates are later inserted into the disambiguation graph, which is used to determine the mapping between entities and mentions. MAG implements two graph-based algorithms to disambiguate entities, i.e., PageRank and HITS. Independently of the chosen graph algorithm, the highest candidate score among the set of candidates is chosen as correct disambiguation for a given mention [4].

3 Demonstration

Our demonstration will show the capabilities of MAG for different languages. We provide a graphical, web-based user interface (GUI). In addition, users can choose to use the REST interface or a Java snippet. For research purposes, MAG can be downloaded and deployed via Maven or Docker. Figure 1 illustrates an example of MAG working on Spanish. The online demo can be accessed via http://agdistis.aksw.org/mag-demo and its code can be downloaded from https://github.com/dice-group/AGDISTIS_DEMO/tree/v2.

Fig. 1.
figure 1

A screenshot of MAG’s web-based demo working on Spanish.

Full size image

We have set up a web service interface for each language version. Each of these interfaces understands two mandatory parameters: (1) text and (2) type.

  1. 1.

    text accepts an UTF-8 and URL encoded string with entities annotated with XML-tag <entity>. It is also capable of recognizing NIF [3] or txt files.

  2. 2.

    type accepts two different values. First, ‘agdistis’ to disambiguate the mentions using the graph-based algorithms, but also ‘candidates’ which list all possible entities for a given mention through the depth-candidate selection of MAG.

Other Parameters. The user can also define more parameters to fine-tune the disambiguation. These parameters have to be set up within the properties fileFootnote 3 or via environment variables while deploying it locally. Below, we describe all the parameters.

  • Popularity - The user can set it as popularity=false or popularity=true. It allows MAG to use either the Page Rank or the frequency of a candidate to sort while candidate retrieval.

  • Graph-based algorithm - The user can choose which graph-based algorithm to use for disambiguating among the candidates per mentions. The current implementation offers HITS and PageRank as algorithms, algorithm=hits or algorithm =pagerank.

  • Search by Context - This boolean parameter provides a search of candidates using a context index [4].

  • Acronyms - This parameter enables a search by acronyms. In this case, MAG uses an additional index to filter the acronyms by expanding their labels and assigns them a high probability. For example, PSG equals Paris Saint-Germain. The parameter is acronym=false or acronym=true.

  • Common Entities - This boolean option supports finding common entities, in case, users desire to find more than ORGANIZATIONs, PLACEs and PERSONs as entity type.

  • Ngram Distance - This integer parameter chooses the ngram distance between words, e.g., bigram, trigram and so on.

  • Depth - This parameter numerically defines how deep the exploration of a semantic disambiguation graph must go.

  • Heuristic Expansion - This boolean parameter defines whether a simple co-occurrence resolution is done or not. For instance, if Barack and Barack Obama are in the same text then Barack is expanded to Barack Obama.

Knowledge-base Agnosticism. Fig. 2 shows a screen capture of our demo for disambiguating mentions using Wikidata. We also provide a web service to allow further investigation. In addition, MAG is used in a domain specific problem using a music Knowledge Base (KB) [5].

Fig. 2.
figure 2

MAG working on Wikidata as Knowledge base.

Full size image

4 Evaluation of the User Interface

We performed a system usability study (SUS)Footnote 4\(^{,}\)Footnote 5 to validate the design of our user interface. 15 users - with a good or no knowledge of Semantic Web, EL or knowledge extraction - selected randomly from all departments at Leipzig University answered our survey. We achieved a SUS-Score of 86.3. This score assigns the mark S to the current interface of MAG and places it into the category of the 10% interfaces, meaning that users of the interface are likely to recommend it to a friend. Figure 3 shows the average voting per question and its standard deviation.

Fig. 3.
figure 3

Standard Usability Score results. The vertical bars show the standard deviation

Full size image

5 Summary

In this demo, we will present MAG, a KB-agnostic and deterministic approach for multilingual EL on 40 different languages contained in DBpedia. Currently, MAG is used in diverse projectsFootnote 6 and has been used largely by the Semantic Web community. We also provide a demo/web-service using Wikidata for supporting an investigation of the graphs structures behind DBpedia and Wikidata pertaining to Information Extraction tasks [1, 2]. The indexes we provided will be used in future work to investigate the EL problem in low-resource languages. Our next step will hence be to evaluate EL on all 40 languages presented in this demo.