Storybase: Towards Building a Knowledge Base for News Events

Proceedings of the AAAI Workshop …, 2006

With the proliferation of news articles from thousands of different sources now available on the Web, summarization of such information is becoming increasingly important. Our research focuses on merging descriptions of news events from multiple sources, to provide a concise description that combines the information from each source. Specifically, we describe and evaluate methods for grouping sentences in news articles that refer to the same event. The key idea is to cluster the sentences, using two novel distance metrics. The first distance metric exploits regularities in the sequential structure of events within a document. The second metric uses a TFIDF-like weighting scheme, enhanced to capture word frequencies within events even though the events themselves are not known a priori. Typical news articles contain sentences that do not describe specific events. We use machine learning methods to differentiate between sentences that describe one or more events, and those that do not. We then remove non-event sentences before initiating the clustering process. We demonstrate that this approach achieves significant improvements in overall clustering performance.

arXiv (Cornell University), 2018

The ever-growing datasets published on Linked Open Data mainly contain encyclopedic information. However, there is a lack of quality structured and semantically annotated datasets extracted from unstructured realtime sources. In this paper, we present principles for developing a knowledge graph of interlinked events using the case study of news headlines published on Twitter which is a real-time and eventful source of fresh information. We represent the essential pipeline containing the required tasks ranging from choosing background data model, event annotation (i.e., event recognition and classification), entity annotation and eventually interlinking events. The state-of-the-art is limited to domain-specific scenarios for recognizing and classifying events, whereas this paper plays the role of a domain-agnostic road-map for developing a knowledge graph of interlinked events.

Event Registry is a system that can analyze news articles and identify in them mentioned world events. The system is able to identify groups of articles that describe the same event. It can identify groups of articles in different languages that describe the same event and represent them as a single event. From articles in each event it can then extract event's core information, such as event location, date, who is involved and what is it about. Extracted information is stored in a database. A user interface is available that allows users to search for events using extensive search options, to visualize and aggregate the search results, to inspect individual events and to identify related events.

Journal of Informetrics, 2007

Really simple syndication (RSS) is becoming a ubiquitous technology for notifying users of new content in frequently updated web sites, such as blogs and news portals. This paper describes a feature-based, local clustering approach for generating overview timelines for major events, such as the tsunami tragedy, from a general-purpose corpus of RSS feeds. In order to identify significant events, we automatically (1) selected a set of significant terms for each day; (2) built a set of (term-co-term) pairs and (3) clustered the pairs in an attempt to group contextually related terms. The clusters were assessed by 10 people, finding that the average percentage apparently representing significant events was 68.6%. Using these clusters, we generated overview timelines for three major events: the tsunami tragedy, the US election and bird flu. The results indicate that our approach is effective in identifying predominantly genuine events, but can only produce partial timelines. (R. Prabowo), Mike.Thelwall@wlv.ac.uk (M. Thelwall), dyner1950@mail.ru (M. Alexandrov).

2013

More and more user-generated content is complementing conventional journalism. While we don't think that CNN or New York Times and its professional journalists will disappear anytime soon, formidable competition is emerging through humble Wikipedia editors. In earlier work (Becker 2012), we found that entertainment and sports news appeared on average about two hours earlier on Wikipedia than on CNN and Reuters online. In this project we build a news-reader that automatically identifies late-breaking news among the most recent Wikipedia articles and then displays it on a dedicated Web site.

Linking Theory and Practice of Digital Libraries, 2021

In this paper, we present an efficient and accurate method to represent events from numerous public sources, such as Wikidata or more specific knowledge bases. We focus on events happening in the real world, such as festivals or assassinations. Our method merges knowledge from Wikidata and Wikipedia article summaries to gather entities involved in events, dates, types and labels. This event characterization procedure is extended by including vernacular languages. Our method is evaluated by a comparative experiment on two datasets that shows that events are represented more accurately and exhaustively with vernacular languages. This can help to extend the research that mainly exploits hub languages, or biggest language editions of Wikipedia. This method and the tool we release will for instance enhance event-centered semantic search engines, a context in which we already use it. An additional contribution of this paper is the public release of the source code of the tool, as well as the corresponding datasets.

Proceedings of the 23rd International Conference on World Wide Web, 2014

Wikipedia's Current Events Portal (WCEP) is a special part of Wikipedia that focuses on daily summaries of news events. The WikiTimes project provides structured access to WCEP by extracting and indexing all its daily news events. In this paper we study this part of Wikipedia and take a closer look into its content and the community behind it. First, we provide descriptive analysis of the collected news events. Second, we compare between the news summaries created by the WCEP crowd and the ones created by professional journalists on the same topics. Finally, we analyze the revision logs of news events over the past 7 years in order to characterize the WCEP crowd and their activities. The results show that WCEP has reached a stable state in terms of the volume of contributions as well as the size of its crowd, which makes it an important source of news summaries for the public and the research community.

International Journal of Computer Applications, 2015

With the advancement in internet technology, users can now easily extract large number of news stories of any ongoing incidents from existing newswires. These hundreds and thousands of news stories causes the problem of information overload. Users find difficulty in capturing the blueprint of the incident as the volume of information is too large. So it becomes necessary to organize news stories into events and learn how these events developed or evolved within the topic. This paper discusses some of the efforts made to model and discover relationships between news events, which had been a less focused area as compared to TDT research which solely focused flat hierarchical structure. A real-world example is discussed for event evolution analysis and future extensions have been proposed.

Zenodo (CERN European Organization for Nuclear Research), 2018

The study presented in this paper is part of our research aimed at improving the search functionalities of current Digital Libraries using formal narratives. Narratives are intended as sequences of events. We present the results of an initial experiment to detect and extract implicit events from the Wikidata knowledge base in order to construct a narrative in a semi-automatic way. Wikidata contains many historical entities, but comparably few events. The reason is that most events in Wikidata are represented in an implicit way, e.g. by listing a date of birth instead of having an event of type "birth". For this reason, we decided to generate what we call the Wikidata Event Graph (WEG), i.e. the graph of implicit events found in Wikidata. We performed an initial experiment taking as case study the narrative of the life of Italian poet Dante Alighieri. Only one event of the life of Dante is explicitly represented in Wikidata as instance of the class Q1190554 Occurrence. Using the WEG, we were able to automatically detect 31 more events of Dante's life that were present in Wikidata in an implicit way.

2009

Abstract News interfaces are largely driven by recent information, even if many events are better interpreted in context of previous events. To address this problem, we consider the task of constructing an explicit representation of a “saga”—a longrunning series of related events. We define a timeline as a concrete representation of a “saga” and we propose two unsupervised methods for timeline construction and compare their performance to hand-produced timelines using a tree edit distance measure.