OTNEL: A Distributed Online Deep Learning Semantic Annotation Methodology

Semantic representation of unstructured text is crucial in modern artificial intelligence and information retrieval applications. The semantic information extraction process from an unstructured text fragment to a corresponding representation from a concept ontology is known as named entity disambiguation. In this work, we introduce a distributed, supervised deep learning methodology employing a long short-term memory-based deep learning architecture model for entity linking with Wikipedia. In the context of a frequently changing online world, we introduce and study the domain of online training named entity disambiguation, featuring on-the-fly adaptation to underlying knowledge changes. Our novel methodology evaluates polysemous anchor mentions with sense compatibility based on thematic segmentation of the Wikipedia knowledge graph representation. We aim at both robust performance and high entity-linking accuracy results. The introduced modeling process efficiently addresses conceptualization, formalization, and computational challenges for the online training entity-linking task. The novel online training concept can be exploited for wider adoption, as it is considerably beneficial for targeted topic, online global context consensus for entity disambiguation.

Cluster-based mention typing for named entity disambiguation

An entity mention in text such as “Washington” may correspond to many different named entities such as the city “Washington D.C.” or the newspaper “Washington Post.” The goal of named entity disambiguation (NED) is to identify the mentioned named entity correctly among all possible candidates. If the type (e.g., location or person) of a mentioned entity can be correctly predicted from the context, it may increase the chance of selecting the right candidate by assigning low probability to the unlikely ones. This paper proposes cluster-based mention typing for NED. The aim of mention typing is to predict the type of a given mention based on its context. Generally, manually curated type taxonomies such as Wikipedia categories are used. We introduce cluster-based mention typing, where named entities are clustered based on their contextual similarities and the cluster ids are assigned as types. The hyperlinked mentions and their context in Wikipedia are used in order to obtain these cluster-based types. Then, mention typing models are trained on these mentions, which have been labeled with their cluster-based types through distant supervision. At the NED phase, first the cluster-based types of a given mention are predicted and then, these types are used as features in a ranking model to select the best entity among the candidates. We represent entities at multiple contextual levels and obtain different clusterings (and thus typing models) based on each level. As each clustering breaks the entity space differently, mention typing based on each clustering discriminates the mention differently. When predictions from all typing models are used together, our system achieves better or comparable results based on randomization tests with respect to the state-of-the-art levels on four defacto test sets.

Good and bad events: combining network-based event detection with sentiment analysis

The huge volume and velocity of media content published on the Web presents a substantial challenge to human analysts. In prior work, we developed a system (network event detection, NED) to assist analysts by detecting events within high-volume news streams in real time. NED can process a heterogeneous stream of news articles or social media user posts, combining text mining and network analysis to detect breaking news stories and generate an easy-to-understand event summary. In this paper, we expand the NED event detection and summarisation approach in two ways. First, we introduce a new approach to named entity disambiguation for tweets, which contain minimal information due to brevity. Second, we apply sentiment analysis techniques to documents associated with a detected event to characterise the event as either broadly ‘positive’ or ‘negative’ based on media portrayal. Our expansion focuses on Twitter streams since Twitter has become an important news dissemination platform and is often the site where emerging events are first seen. To test the extended methodology, we apply it here to three data sets related to political elections in the UK and the USA. The addition of sentiment analysis to the NED event detection methodology improves the insight gained by the user by allowing quick evaluation of the perceived impact of an event. This approach may have potential applications in domains where public sentiment is relevant to decision-making around events, such as financial markets and politics.

Analysis of Name Entities in Text Using Robust Disambiguation Method

<em>Named entities are proper nouns or objects contained in a text, such as a person's name, country name, and others. Names of persons in some text are often ambiguous, which makes it difficult for ordinary people to find out these same names are the same person or not. An ambiguity of names also found in hadith, like the name Abdullah in hadith number 86 and 2411, that might be the same person or might be different. Based on this problem, then this study focuses on named entity disambiguation, which considered further semantic and lexical relation between a named entity. Expected in the future, it would help people to understand the ambiguity of the name or distinguish ambiguous names. The method used in this research was Robust Disambiguation because, in this method, the context of the named entity considered. The resulted output obtained was in the form of named entity that grouped based on the same person or different person processed with Density-based Spatial Clustering of Applications with Noise. This research resulted in an accuracy value of 90%, a precision value of 97%, and a recall value of 89% obtained from actual value and predicted value</em>