Luka Bradesko

OntoPlus : text-driven ontology extension using ontology content, structure and co-occurrence information

This paper addresses the process of semi-automatic text-driven ontology extension using ontology content, structure and co-occurrence information. A novel OntoPlus methodology is proposed for semi-automatic ontology extension based on text mining methods. It allows for the effective extension of the large ontologies, providing a ranked list of potentially relevant concepts and relationships given a new concept (e.g., glossary term) to be inserted in the ontology. A number of experiments are conducted, evaluating measures for ranking correspondence between existing ontology concepts and new domain concepts suggested for the ontology extension. Measures for ranking are based on incorporating ontology content, structure and co-occurrence information. The experiments are performed using a well known Cyc ontology and textual material from two domains – finances and, fisheries & aquaculture. Our experiments show that the best results are achieved by combining content, structure and co-occurrence information. Furthermore, ontology content and structure seem to be more important than co-occurrence for our data in the financial domain. At the same time, ontology content and co-occurrence seem to have higher importance for our fisheries & aquaculture domain.

Curious Cat--Mobile, Context-Aware Conversational Crowdsourcing Knowledge Acquisition

Scaled acquisition of high-quality structured knowledge has been a longstanding goal of Artificial Intelligence research. Recent advances in crowdsourcing, the sheer number of Internet and mobile users, and the commercial availability of supporting platforms offer new tools for knowledge acquisition. This article applies context-aware knowledge acquisition that simultaneously satisfies users’ immediate information needs while extending its own knowledge using crowdsourcing. The focus is on knowledge acquisition on a mobile device, which makes the approach practical and scalable; in this context, we propose and implement a new KA approach that exploits an existing knowledge base to drive the KA process, communicate with the right people, and check for consistency of the user-provided answers. We tested the viability of the approach in experiments using our platform with real users around the world, and an existing large source of common-sense background knowledge. These experiments show that the approach is promising: the knowledge is estimated to be true and useful for users 95% of the time. Using context to proactively drive knowledge acquisition increased engagement and effectiveness (the number of new assertions/day/user increased for 175%). Using pre-existing and newly acquired knowledge also proved beneficial.

Curious cat conversational crowd based and context aware knowledge acquisition chat bot

Acquisition of high quality structured knowledge that is immediately useful for reasoning algorithms has been a longstanding goal of the Artificial Intelligence research community. With the recent advances in crowdsourcing, the sheer number of internet users and the commercial availability of supporting platforms have come a new set of tools to tackle this problem. Although numerous systems and methods for crowdsourced knowledge acquisition had been developed and solve the problem of manpower, the issues of task preparation, financial cost, finding the right crowd, consistency, and quality of the acquired knowledge, seem to persist. In this paper we propose a new approach to address this deficit by exploiting an existing knowledge base to drive the acquisition process, address the right people, and check their answers for consistency. We conducted tests of the viability of the approach in experiments with real users, a working platform and common sense knowledge.

Isaac Bloomberg Meets Michael Bloomberg: Better EntityDisambiguation for the News

This paper shows the implementation and evaluation of the Entity Linking or Named Entity Disambiguation system used and developed at Bloomberg. In particular, we present and evaluate a methodology and a system that do not require the use of Wikipedia as a knowledge base or training corpus. We present how we built features for disambiguation algorithms from the Bloomberg News corpus, and how we employed them for both single-entity and joint-entity disambiguation into a Bloomberg proprietary knowledge base of people and companies. Experimental results show high quality in the disambiguation of the available annotated corpus.

Supporting Rule Generation and Validation on Environmental Data in EnStreaM

Detection rules represent one of the components of the rule models in event processing systems. These rules can be discovered from data using data mining techniques or domain experts’ knowledge. We demonstrate a system that provides its users the means for creating and validating such rules. The system is applied on real-life environmental scenarios, where the main source of data comes from sensors. Based on historical data about events of interest, the scope is to formulate rules that could have caused these events. Using a scalable infrastructure the rules can be tested on massive amount of data in order to observe how past events would fit to these rules. In addition, we create semantic annotations of the dataset and use them in the system outputs in order to support interoperability with other systems.

From mobility patterns to behavioural change: leveraging travel behaviour and personality profiles to nudge for sustainable transportation

Rendering transport behaviours more sustainable is a pressing issue of our times. In this paper, we rely on the deep penetration of mobile phones in order to influence citizens’ behavior through data-driven mobility and persuasive profiles. Our proposed approach aims to nudge users on a personalized level in order to change their mobility behavior and make more sustainable choices. To achieve our goal, first we leverage pervasive mobile sensing to uncover users’ mobility patterns and use of transportation modes. Second, we construct users’ persuadability profiles by considering their personality and mobility behavior. With the use of the aforementioned information we generate personalized interventions that nudge users to adopt sustainable transportation habits. These interventions rely on persuasive technologies and are embedded in a route planning application for smartphones. A pilot study with 30 participants using the system for 6 weeks provided fairly positive evaluation results in terms of the acceptance of our approach and revealed instances of behavioural change.

Estimating point-of-interest rating based on visitors geospatial behaviour

Rating of different services, products and experiences plays an important role in our digitally assisted day-to-day life. It helps us make decisions when we are indecisive, uninformed or inexperienced. Traditionally, ratings depend on the willingness of existing customers to provide them. This often leads to biased (due to the insufficient number of votes) or nonexistent ratings. This was the motivation for our research, which aims to provide automatic star rating. The paper presents an approach to extracting points-of-interest from various sources and a novel approach to estimating point-of-interest ratings, based on geospatial data of their visitors. Our research is applied to campsite dataset where the community is still developing and more than thirty percent of camps are unrated. Our study use case addresses a realword problem of motorhome users visiting campsites in European countries. The dataset includes GPS traces from 10 motorhomes that were collected over a period of 2 years. To estimate star ratings of points-of-interest we applied machine learning methods including support vector machine, linear regression, random forest and decision trees. Our experimental results show that the duration of visit, which is a crucial part of the proposed approach, is an indicative feature for predicting camp ratings.