MultiWiki

Question Answering (QA) systems play an important role in decision support systems. Deep neural network-based passage rankers have recently been developed to more effectively rank likely answer-containing passages for QA purposes. These rankers utilize distributed word or sentence embeddings. Such distributed representations mostly carry semantic relatedness of text units in which explicit linguistic features are under-represented. In this paper, we take novel approaches to combine linguistic features (such as different part-of-speech measures) with distributed sentence representations of questions and passages. The QUASAR-T fact-seeking questions and short text passages were used in our experiments to show that while ensembling of deep relevance measures based on pure sentence embedding with linguistic features using several machine learning techniques fails to improve upon the passage ranking performance of our baseline neural network ranker, the concatenation of the same features within the network structure significantly improves the overall performance of passage ranking for QA.

At the New York University College of Dentistry, lab conference hours for the review of anatomy in preparation for the National Board Dental Examination Part I (NBDE Part I) were reduced due to an increase in incoming class size in 2015. Tasked with providing students an effective means to review material that could now not be covered in the given lab hours, we identified the online platform Cerego as a way to provide a self‐guided anatomy study tool. Use of Cerego replaced what would otherwise have been 96 faculty hours of instruction.We initially created three anatomy topic modules in Cerego, each of which included a combination of images, identifications, associations, and short text passages. Creation of modules in Cerego was intuitive and straightforward. The platform, with algorithms based on principles of retrieval learning and spaced practice, guides the student through individualized iterations of review in multiple formats, and provides the student with prompts for the most effective timing for return to the material. In addition, both student and instructor have access to fine grained performance analytics.We found that Cerego is well suited to fact memorization, and effective as a means to learn material within the time‐frame of a single semester. However, full mastery using the Cerego program – leading to long‐term retention – is achieved over a more extended period. A survey after the termination of the NBDE Part I course in 2015, as well as informal discussions with students, has indicated high student satisfaction with Cerego as a study tool. In addition, results for the anatomy portion of the board exam demonstrates student success, with NYU Dentistry students as a whole achieving a 97.1–100% first‐time pass rate, and performing at 1.4–2.6 standard deviations above the mean for all dental schools in anatomy in the years 2015–2017. Given the success of our initial Cerego use, we have now incorporated this platform into the Head and Neck Anatomy and Basic Tissues courses within the Department of Basic Science and Craniofacial Biology, and other departments at NYU Dentistry have also adopted the use of this tool.This abstract is from the Experimental Biology 2018 Meeting. There is no full text article associated with this abstract published in The FASEB Journal.

<p>We describe the investigation of automatic annotation of text with pictures, where knowledge extraction uses dependency parsing. Annotation of text with pictures, a form of knowledge visualization, can assist understanding. The problem statement is, given a corpus of images and a short passage of text, extract knowledge (or concepts), and then display that knowledge in pictures along with the text to help with understanding. A proposed solution framework includes a component to extract document concepts, a component to match document concepts with picture metadata, and a component to produce an amalgamated output of text and pictures. A proof-of-concept application based on the proposed framework provides encouraging results</p>

The rapid growth of scientific papers makes it difficult to find relevant and appropriate citations. Context-aware citation recommendation aims to overcome this problem by providing a list of scientific papers given a short passage of text. In this paper, we propose a long-short-term memory (LSTM)based model for context-aware citation recommendation, which first learns the distributed representations of the citation contexts and the scientific papers separately based on LSTM, and then measures the relevance based on the learned distributed representation of citation contexts and the scientific papers. Finally, the scientific papers with high relevance scores are selected as the recommendation list. In particular, we try to incorporate author information, venue information, and content information in scientific paper distributed vector representation. Furthermore, we integrate author information of the given context in citation context distributed vector representation. Thus, the proposed model makes personalized context-aware citation recommendation possible, which is a new issue that few papers addressed in the past. When conducting experiments on the ACL Anthology Network and DBLP data sets, the results demonstrate the proposed LSTM-based model for context-aware citation recommendation is able to achieve considerable improvement over previous context-aware citation recommendation approaches. The personalized recommendation approach is also competitive with the non-personalized recommendation approach.

The English language Wikipedia is notable for its large number of articles. However, 288 other active language editions of Wikipedia have also developed through the intricate interactions of contributing editors. While the editor interactions in the English Wikipedia have been researched extensively, these other language editions remain understudied. To understand how editors currently come to consensus in article building in the Spanish language, a team of researchers has leveraged an existing English framework that depicts how power and policies play a role in mass collaboration. Using this English language framework, we are utilizing qualitative coding methods to build a unique model of the editor interactions on the Spanish language Wikipedia (ES). The results of this study will help contribute to a deeper understanding of how a framework in a different language edition of Wikipedia varies from the English. Our preliminary findings show that the power plays utilized in the Spanish Wikipedia talk pages differ in type and amount from power plays identified in the talk pages in the English language Wikipedia (EN), suggesting different forms of editor interactions and facilitation strategies across this multilingual platform.

In this paper we study the differences in historical worldviews between Western and Eastern cultures, represented through the English, Chinese and Japanese Wikipedia. In particular, we analyze the historical networks of the world’s leaders since the beginning of written history, comparing them in the three different language versions of Wikipedia.

The advent of blogging more and more content is published on the web. Micro blogging is used to convey ideas using short text messages; video links or images .micro blogging platforms allow the public to convey their ideas through new technologies and to convey ideas and news in precise manner. This paper aims to use to text mining algorithm, which is used for gathering relevant information from passages of text, to derive a context through which predictive analysis can be made. The goal will be to use this technique to provide a predictive analysis on the electoral polls using tweets from the micro-blogging platform twitter.

This paper proposes a neural network-based method for generating compact answers to open-domain why-questions (e.g., "Why was Mr. Trump elected as the president of the US?"). Unlike factoid question answering methods that provide short text spans as answers, existing work for why-question answering have aimed at answering questions by retrieving relatively long text passages, each of which often consists of several sentences, from a text archive. While the actual answer to a why-question may be expressed over several consecutive sentences, these often contain redundant and/or unrelated parts. Such answers would not be suitable for spoken dialog systems and smart speakers such as Amazon Echo, which receive much attention in these days. In this work, we aim at generating non-redundant compact answers to why-questions from answer passages retrieved from a very large web data corpora (4 billion web pages) by an already existing open-domain why-question answering system, using a novel neural network obtained by extending existing summarization methods. We also automatically generate training data using a large number of causal relations automatically extracted from 4 billion web pages by an existing supervised causality recognizer. The data is used to train our neural network, together with manually created training data. Through a series of experiments, we show that both our novel neural network and auto-generated training data improve the quality of the generated answers both in ROUGE score and in a subjective evaluation.

A multi-view similarity measure framework for trouble ticket mining

Human language is naturally fuzzy by nature, with words meaning different things to different people, depending on the context. Fuzzy words are words with a subjective meaning, which are typically used in everyday human natural language dialogue and are often ambiguous and vague in meaning and are based on an individual’s perception. Fuzzy Sentence Similarity Measures (FSSM) are algorithms that can compare two or more short texts which contain human-perception-based words and return a numeric measure of similarity of meaning between them. This paper proposes a new FSSM called FUSE (FUzzy Similarity mEasure), to assess an individual’s perception within a FSSM. FUSE is an ontology-based similarity measure that uses Interval Type-2 fuzzy sets to model relationships between categories of human perception-based words. The FUSE algorithm has been developed over four versions and evaluated on several published and newly created datasets. Typically, results have shown that calculating the semantic similarity of two short texts using FUSE, gives a higher correlation with the average human ratings (AHR) compared to traditional sentence similarity measures that do not consider the presence of fuzzy words. This paper focuses on the second version of the FUSE algorithm, referred to as FUSE_2.0 which has been compared to several state-of-the-art, semantic similarity measures (SSM), including the only published FSSM, FAST (Fuzzy Algorithm for Similarity Testing), which has a limited dictionary of fuzzy words and uses Type-1 to model relationships between categories of human perception-based words. Results have shown that FUSE_2.0 achieves a higher correlation with the average human ratings (AHR) compared to traditional SSM’s and FAST. The key contributions of this work can be summarised as follows: The development of a new methodology to model fuzzy words using Interval Type-2 fuzzy sets. This has led to the creation of a fuzzy dictionary for nine fuzzy categories, a useful resource which can be used by other researchers in the field of natural language processing and Computing with Words (CWW) with other fuzzy applications such as semantic clustering.

Accurate alignment of the watch hand is a critical procedure in the industrial production of watches. However, compared with the traditional ideal laboratorial case, the watch-hand alignment suffers various external disturbances from the aged equipment and industrial operating conditions, which seriously influence the alignment precision. To achieve accurate watch-hand alignment in the complex industrial environment, this paper develops a robotic micromanipulation system and proposes a corresponding robust control strategy. First, a micromanipulation system with five degrees of freedom is set up and integrated with an optical microscope. Then, a set of proper machine vision methods is adopted to obtain accurate position information in disregard of the interference terms. Third, the dynamic model of the controlled watch hand is built with consideration of the external disturbances from the production environment. Since the velocity information cannot be observed accurately through the machine vision process, we design a state observer to acquire the position and velocity values for the following control strategy. After that, a sliding mode controller with the radical basis function neural network is proposed to achieve high-precision alignment while resisting the external disturbance. Finally, simulation and experimental results prove that the proposed method can realize the alignment of the watch hand within 6 s with the accuracy of $2\times 10^{-3}$ rad, which is enhanced at least two times compared with the traditional manual method. Note to Practitioners —This paper was motivated by the challenging problem of the watch-hand alignment in the industrial production. Unlike the traditional manipulation, this alignment is subjected to various disturbances from the aged equipment and complex operating conditions, including complex forces at the air–liquid operation interface and the confusing scratches, undesired flaking particles, as well as the reflection of light in the image. Here, we propose a microrobotic system and a robust control strategy to precisely align the watch hand and eliminate the influence from the above disturbances. In the machine vision process for recognizing the object, filtering, contour/corner detection, and template matching methods are adopted to acquire accurate position information despite the interference in image processing. To avoid the disturbance from the complex forces during the alignment, we develop the robust sliding mode controller to guarantee the alignment precision. The experiment results indicate that the proposed system can align the watch hand with the accuracy enhanced at least two times as previous manual method. This paper provides a general solution for micromanipulation under disturbance in industrial production environment.

User acceptance of new technologies is a key aspect to implement successful human-robot teams at work. The acceptance of new technologies, according to the Unified theory of Acceptance and Use Model, is determined by user's behavioral intention to use a technology and based on performance expectancy, effort expectancy, social influence and facilitating conditions. Furthermore, it is influenced by the experience a user has made with the new technology in the past. As not everyone can make their own experiences with robots in the workplace, many develop a first opinion by media consumption. To get an overview on media opinions on robots in working contexts, we extracted and coded the tone of headlines of three major English online newspapers. Subsequently, we compared the results of a human coding team with an open source approach for sentiment analysis. The comparison between human coders and the algorithm showed diverging results. Based on our analysis a team of human coders seems to be better in tone coding of short text passages like headlines. The overall results showed a negative overall tone in online newspaper headlines from 2016 to 2018. While 24% of the identified headlines had a positive tone, 43% contained a negative tone and 33% of the included headlines were coded as neutral. The negative tone in headlines on robots at work and their possible impact on workers expectations should be taken into account when introducing these new technologies in work contexts to ensure a human-centered implementation.

Wikipedia, in its role as the world's largest encyclopedia, serves a broad range of information needs. Although previous studies have noted that Wikipedia users' information needs vary throughout the day, there is to date no large-scale, quantitative study of the underlying dynamics. The present paper fills this gap by investigating temporal regularities in daily consumption patterns in a large-scale analysis of billions of timezone-corrected page requests mined from English Wikipedia's server logs, with the goal of investigating how context and time relate to the kind of information consumed. First, we show that even after removing the global pattern of day-night alternation, the consumption habits of individual articles maintain strong diurnal regularities. Then, we characterize the prototypical shapes of consumption patterns, finding a particularly strong distinction between articles preferred during the evening/night and articles preferred during working hours. Finally, we investigate topical and contextual correlates of Wikipedia articles' access rhythms, finding that article topic, reader country, and access device (mobile vs. desktop) are all important predictors of daily attention patterns. These findings shed new light on how humans seek information on the Web by focusing on Wikipedia as one of the largest open platforms for knowledge and learning, emphasizing Wikipedia's role as a rich knowledge base that fulfills information needs spread throughout the day, with implications for understanding information seeking across the globe and for designing appropriate information systems.

This essay uncovers the systematic, intentional distortion of Holocaust history on the English-language Wikipedia, the world’s largest encyclopedia. In the last decade, a group of committed Wikipedia editors have been promoting a skewed version of history on Wikipedia, one touted by right-wing Polish nationalists, which whitewashes the role of Polish society in the Holocaust and bolsters stereotypes about Jews. Due to this group’s zealous handiwork, Wikipedia’s articles on the Holocaust in Poland minimize Polish antisemitism, exaggerate the Poles’ role in saving Jews, insinuate that most Jews supported Communism and conspired with Communists to betray Poles (Żydokomuna or Judeo–Bolshevism), blame Jews for their own persecution, and inflate Jewish collaboration with the Nazis. To explain how distortionist editors have succeeded in imposing this narrative, despite the efforts of opposing editors to correct it, we employ an innovative methodology. We examine 25 public-facing Wikipedia articles and nearly 300 of Wikipedia’s back pages, including talk pages, noticeboards, and arbitration cases. We complement these with interviews of editors in the field and statistical data gleaned through Wikipedia’s tool suites. This essay contributes to the study of Holocaust memory, revealing the digital mechanisms by which ideological zeal, prejudice, and bias trump reason and historical accuracy. More broadly, we break new ground in the field of the digital humanities, modelling an in-depth examination of how Wikipedia editors negotiate and manufacture information for the rest of the world to consume.

<p><strong>Background</strong></p><p>Wikipedia is an open source, web-based encyclopedia, allowing anonymous and registered users to create, edit and improve articles. A survey in 2018 showed that as many as 90% of Wikipedia's editors were male and as many as 81% of contributors were from the Global North [1].  </p><p>In addition, there are fewer contributions about women, especially in STEM fields, and they are usually less developed [2]. In October 2014, only 15.53% of English Wikipedia's biographies were about women [3]. The WikiProject <strong>Women in Red</strong> was founded in July 2015 with the objective to address this gender bias in Wikipedia content. They succeeded in increasing the above-mentioned percentage to <strong>18.71%</strong> as of 11 January 2021 [2].</p><p>Today, Wikipedia is within the 20 most popular websites [4] and every month it attracts more than 1 billion unique visitors [5]. Wikipedia therefore has a huge potential to change publics perception of who is doing science and what a scientist ‘looks’ like.</p><p><strong>(Women) planetary scientists on Wikipedia?</strong></p><p>In June 2020, there were only 189 planetary scientist biographies on the English Wikipedia, including 48 biographies of female planetary scientists (25%). This percentage is in agreement with the percentage of women in the International Astronomical Union from all ESA’s Member States (24%) [6], but planetary scientists are clearly underrepresented on Wikipedia. Many of them either do not have a Wikipedia biography yet, or if they do, they are often misclassified under the category of “astronomers” or “astrophysicists”.</p><p><strong>A Planetary Sciences</strong><strong> Edit-a-thon</strong>    </p><p>The Diversity Committee of the Europlanet Society aims to highlight diversity within the planetary science community. Therefore, they organised, in collaboration with Women in Red and WikiDonne, the first <strong>Planetary Science Wiki Edit-a-thon</strong> during the Europlanet Science Congress (EPSC) 2020 [7]. An Edit-a-thon (‘edit marathon’) is an organized event where editors from an online community (such as Wikipedia in this case) write, translate and improve articles on a specific topic [2]. Thirty persons received a basic editing training, resulting in 1 new article and 5 translated ones. A small subgroup still meets every month to continue the project.</p><p><strong>References</strong></p><p>[1] https://meta.wikimedia.org/wiki/Community_Insights/2018_Report</p><p>[2] https://en.wikipedia.org/wiki/Wikipedia:WikiProject_Women_in_Red  </p><p>[3] Eduardo Graells-Garrido, Mounia Lalmas, Filippo Menczer, "First Women, Second Sex: Gender Bias in Wikipedia", arXiv, 9 February 2015, p. 3.</p><p>[4] "Wikipedia.org Traffic, Demographics and Competitors". www.alexa.com. Retrieved October 1, 2019.</p><p>[5] https://stats.wikimedia.org/#/all-wikipedia-projects</p><p>[6] Piccialli A., et al., Participation of women scientists in ESA Solar System missions: a historical trend, Adv. Geosci., 53, 169–182, https://doi.org/10.5194/adgeo-53-169-2020, 2020.</p><p>[7] https://www.europlanet-society.org/epsc-2020-planetary-science-wiki-edit-a-thon/    </p>