The massive shock of the COVID-19 pandemic has already shown its negative effects on economies around world, unprecedented in recent history. infections and containment measures caused a general slowdown research new knowledge production. Because link between R&D output economic growth, it is to be expected then that activities will slow turn global recovery from pandemic. Many studies also claim an uneven impact scientific production across gender. In this paper, we investigate phenomenon...

In a conversational context, user expresses her multi-faceted information need as sequence of natural-language questions, i.e., utterances. Starting from given topic, the conversation evolves through utterances and system replies. The retrieval documents relevant to utterance in is challenging due ambiguity natural language difficulty detecting possible topic shifts semantic relationships among We adopt 2019 TREC Conversational Assistant Track (CAsT) framework experiment with modular...

Web usage mining is the application of data techniques to generated by interactions users with web servers. This kind data, stored in server logs, represents a valuable source information, which can be exploited optimize document-retrieval task, or better understand, and thus, satisfy user needs.

Tracking public opinion in social media provides important information to enterprises or governments during a decision making process. In addition, identifying and extracting the causes of sentiment spikes allows interested parties redesign adjust strategies with aim attract more positive sentiments. this paper, we focus on problem tracking towards different entities, detecting ranking spike. Our approach combines LDA topic model Relative Entropy. The former is used for topics discussed time...

Privacy of Internet users is at stake because they expose personal information in posts created online communities, search queries, and other activities. An adversary that monitors a community may identify the with most sensitive properties utilize this knowledge against them (e.g., by adjusting pricing goods or targeting ads nature). Existing privacy models for structured data are inadequate to capture risks from user posts.

Suggesting personalized venues helps users to find interesting places on location-based social networks (LBSNs). Although there are many LBSNs online, none of them is known have thorough information about all venues. The Contextual Suggestion track at TREC aimed providing a collection consisting as well user context enable researchers examine and compare different approaches, under the same evaluation setting. However, officially released did not meet participants' needs related venue...

We design algorithms that, given a collection of documents and distribution over user queries, return small subset the document in such way that we can efficiently provide high-quality answers to queries using only selected subset. This approach has applications when space is constraint or query-processing time increases significantly with size collection. study our through lens stochastic analysis prove even though they use fraction entire collection, most achieving performance close...

In this paper we present a novel graph-based data abstraction for modeling the browsing behavior of web users. The objective is to identify users who discover interesting pages before others. We call these early adopters. By tracking activity adopters can new early, and recommend similar focus on news blog pages, which are more dynamic in nature appropriate recommendation.

With the rapid proliferation of microblogging services such as Twitter, a large number tweets is published everyday often making users feel overwhelmed with information. Helping these to discover potentially interesting an important task for services. In this paper, we present novel tweet-recommendation approach, which exploits network, content, and retweet analyses recommendations tweets. The idea recommend that are not visible user (i.e., they do appear in timeline) because nobody her...

This paper proposes a new model of user-centric, global, probabilistic privacy, geared for today's challenges helping users to manage their privacy-sensitive information across wide variety social networks, online communities, QA forums, and search histories. Our approach anticipates an adversary that harnesses global background knowledge rich statistics in order make educated guesses, is, inferences at sensitive data. We aim tool simulates such powerful adversary, predicts privacy risks,...

Crowdsourcing is a computational paradigm whose distinctive feature the involvement of human workers in key steps computation. It used successfully to address problems that would be hard or impossible solve for machines. As we highlight this work, exclusive use nonexpert individuals may prove ineffective some cases, especially when task at hand need accurate solutions demand degree specialization avoid excessive uncertainty and inconsistency answers. We limitation by proposing an approach...

Making personalized and context-aware suggestions of venues to the users is very crucial in venue recommendation. These are often based on matching venues' features with users' preferences, which can be collected from previously visited locations. In this paper we present a novel user-modeling approach relies set scoring functions for making content reviews as well context. Our experiments, conducted dataset TREC Contextual Suggestion Track, proved that our methodology outperforms...

Abstract In this paper, we address the problem of answering complex questions formulated by users in natural language. Since traditional information retrieval systems are not suitable for questions, these usually run over knowledge bases, such as Wikidata or DBpedia. We propose a semi-automatic approach transforming language question into SPARQL query that can be easily processed base. The applies classification techniques to associate with proper template from set predefined templates....

In this paper we introduce the problem of query covering as a means to efficiently cache results. The general idea is populate with documents that contribute result pages large number queries, opposed caching top for each query. It turns out hard and solving it requires knowledge structure queries results space, well input distribution. We formulate under framework stochastic optimization; theoretically can be seen universal version set multicover. While NP-hard solved exactly, show any...

Topic modeling is an important area which aims at indexing and exploring massive data streams. In this paper we introduce a discrete Dynamic Modeling (dDTM) algorithm, able to model dynamic topic that not necessarily present over all time slices in stream of documents. Our proposed has applications topics rapidly changing less structured data, such as online microblogs news

Nowadays, on-line news agents post articles on social media platforms with the aim to spread information as well attract more users and understand their reactions opinions. Predicting emotional influence of is very important not only for but also users, who can filter out based they trigger. In this paper, we focus problem prediction a before publication. For prediction, explore range textual semantic features derived from content posts. Our results show that terms most feature extracted...