Entity linking by focusing DBpedia candidate entities

Abstract

Recently, Entity Linking and Retrieval turned out to be one of the most interesting tasks in Information Extraction due to its various applications. Entity Linking (EL) is the task of detecting mentioned entities in a text and linking them to the corresponding entries of a Knowledge Base. EL is traditionally composed of three major parts: i)spotting, ii)candidate generation, and iii)candidate disambiguation. The performance of an EL system is highly dependent on the accuracy of each individual part. In this paper, we focus on these three main building blocks of EL systems and try to improve on the results of one of the open source EL systems, namely DBpedia Spotlight. We propose to use text pre-processing and parameter tuning to "focus" a general-purpose EL system to perform better on different kinds of input text. Also, one of the main drawbacks of EL systems is identifying where a name does not refer to any known entity. To improve this so-called NIL-detection, we define different features using a set of texts and their known entities and design a classifier to automatically classify DBpedia Spotlight's output entities as "NIL" or "Not NIL". The proposed system has participated in the SIGIR ERD Challenge 2014 and the performance analysis of this system on the challenge's datasets shows that the proposed approaches successfully improve the accuracy of the baseline system.

Scalable Disambiguation System Capturing Individualities of Mentions

“…Supervised global ranking models: On the other hand, non-graph-based solutions [4,9,15,16,17,19] are mostly supervised in the linking phase. Milne and Witten [17] assumed that there exists unambiguous mentions associated with a single sense, and evaluated the relatedness between candidate senses and unambiguous mentions (senses).…”

“…Han and Sun [9] proposed a generative probabilistic model, using the frequency of mentions and context words given a candidate sense, as independent generative features; this statistical model is also the core module of the public disambiguation service DBpedia Spotlight [4]. Olieman et al [19] proposed various adjustments (calibrating parameters, preprocessing text input, merging normal and capitalized results) to adapt Spotlight to both short and long texts. They also used a global binary classifier with several similarity metrics to prune off uncertain Spotlight results.…”

Improving Language-Dependent Named Entity Detection

“… POS tags and rules. A couple of authors use part of speech (POS) taggers and/or several rules in order to identify named entities [19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34]. The rules range from simple rules such as "capitalized letter" (if a word contains a capitalized letter the word will be treated as a spot), stop word lists, "At Least One Noun Selector"-rule to complex, combined rules.…”

“…Other authors use N-Grams [31,33,44,51,52], others experiment with CRF [25,26], Topic Modeling [53,54], Naïve Bayes and Hidden Markov Models [55]. Last but not least one can find approaches based on Finite-state machines [14,15,30,34].…”

Erd'14

“…The second speaker was Jaap Kamps from the University of Amsterdam who presented the Expose system [21] for ERD. The focus of this team was to improve on the results of an open source ERD system, namely DBpedia Spotlight.…”