Formal language models for finding groups of experts

The task of finding groups or teams has recently received increased attention, as a natural and challenging extension of search tasks aimed at retrieving individual entities. We introduce a new group finding task: given a query topic, we try to find knowledgeable groups that have expertise on that topic. We present five general strategies for this group finding task, given a heterogenous document repository. The models are formalized using generative language models. Two of the models aggregate expertise scores of the experts in the same group for the task, one locates documents associated with experts in the group and then determines how closely the documents are associated with the topic, whilst the remaining two models directly estimate the degree to which a group is a knowledgeable group for a given topic. For evaluation purposes we construct a test collection based on the TREC 2005 and 2006 Enterprise collections, and define three types of ground truth for our task. Experimental results show that our five knowledgeable group finding models achieve high absolute scores. We also find significant differences between different ways of estimating the association between a topic and a group.     

Studying health anxiety related attentional bias during online health information seeking: Impacts of stages and task types

Seeking online health information may reinforce the anxiety of those who are already overly anxious about their health. This study explored how people with health anxiety may behave differently in terms of their attentional biases when seeking health information online. We conducted an eye-tracking experiment with 17 participants in the high health anxious group and 17 participants in the low health anxious group, who performed three types of information-seeking tasks (factual, interpretive, and exploratory) on a Chinese health website. We observed that both groups mainly allocated their attention to the stages of evaluating the list of search results and synthesizing information to make health decisions. They showed similar attention tracks at the earlier search stages and health anxiety was found to associate with attentional biases towards certain website stimuli. However, the high health anxious group showed more active eye movements than their low health anxious counterparts. Attentional biases from the high health anxious group mainly occurred at the later stage of processing rather than the initial orientation stages. As for task types, the high health anxious group presented more extensive attentional biases when performing the interpretive task, compared to the explorative and factual tasks. The findings provide novel insights into the attentional biases of people with health anxiety as they search online for health information, which have implications on designing more effective information interventions for vulnerable groups of health information consumers. The findings can also help clinicians interpret patients’ anxiety-related sensations and provide intervening recommendations for clients in use of online health information.     

Beyond the search process – Exploring group members’ information behavior in context

This paper reports on the findings from a longitudinal case study exploring Kuhlthau’s information search process (ISP)-model in a group based academic setting. The research focus is on group members’ activities and cognitive and emotional experiences during the task process of writing an assignment. It is investigated if group members’ information behavior differ from the individual information seeker in the ISP-model and to what extent this behavior is influenced by contextual (work task) and social (group work) factors. Three groups of LIS students were followed during a 14 weeks period in 2004/2005 (10 participants). Quantitative and qualitative methods were employed, such as demographic surveys, process surveys, diaries and interviews. Similarities in behavior were found between group members and the individual in Kuhlthau’s ISP-model with regard to the general stages of information seeking, the cognitive pattern associated with focus formulation and the tendency towards an increase in writing activities while searching activities decreased. Differences in behavior were also found, which were associated with contextual and social factors beyond the mere search process. It is concluded that the ISP-model does not fully comply with group members’ problem solving process and the involved information seeking behavior. Further, complex academic problem solving seems to be even more complex when it is performed in a group based setting. The study contributes with a new conceptual understanding of students’ behavior in small groups.     

Emotional reactions to robot colleagues in a role-playing experiment

We investigated how people react emotionally to working with robots in three scenario-based role-playing survey experiments collected in 2019 and 2020 from the United States (Study 1: N = 1003; Study 2: N = 969, Study 3: N = 1059). Participants were randomly assigned to groups and asked to write a short post about a scenario in which we manipulated the number of robot teammates or the size of the social group (work team vs. organization). Emotional content of the corpora was measured using six sentiment analysis tools, and socio-demographic and other factors were assessed through survey questions and LIWC lexicons and further analyzed in Study 4. The results showed that people are less enthusiastic about working with robots than with humans. Our findings suggest these more negative reactions stem from feelings of oddity in an unusual situation and the lack of social interaction.     

How doctors search: A study of query behaviour and the impact on search results

Professional, workplace searching is different from general searching, because it is typically limited to specific facets and targeted to a single answer. We have developed the semantic component (SC) model, which is a search feature that allows searchers to structure and specify the search to context-specific aspects of the main topic of the documents. We have tested the model in an interactive searching study with family doctors with the purpose to explore doctors’ querying behaviour, how they applied the means for specifying a search, and how these features contributed to the search outcome. In general, the doctors were capable of exploiting system features and search tactics during the searching. Most searchers produced well-structured queries that contained appropriate search facets. When searches failed it was not due to query structure or query length. Failures were mostly caused by the well-known vocabulary problem. The problem was exacerbated by using certain filters as Boolean filters. The best working queries were structured into 2–3 main facets out of 3–5 possible search facets, and expressed with terms reflecting the focal view of the search task. The findings at the same time support and extend previous results about query structure and exhaustivity showing the importance of selecting central search facets and express them from the perspective of search task. The SC model was applied in the highest performing queries except one. The findings suggest that the model might be a helpful feature to structure queries into central, appropriate facets, and in returning highly relevant documents.     

Searching for Safety Online: Managing "Trolling" in a Feminist Forum

A common phenomenon in online discussion groups is the individual who baits and provokes other group members, often with the result of drawing them into fruitless argument and diverting attention from the stated purposes of the group. This study documents a case in which the members of an online community--a feminist web-based discussion forum--are targeted by a "troll" attempting to disrupt their discussion space. We analyze the strategies that make the troller successful and the targeted group largely ineffectual in responding to his attack, as a means to understand how such behavior might be minimized and managed in general. The analysis further suggests that feminist and other nonmainstream online forums are especially vulnerable, in that they must balance inclusive ideals against the need for protection and safety, a tension that can be exploited by disruptive elements to generate intragroup conflict.     

Text classification from unlabeled documents with bootstrapping and feature projection techniques

Many machine learning algorithms have been applied to text classification tasks. In the machine learning paradigm, a general inductive process automatically builds a text classifier by learning, generally known as supervised learning. However, the supervised learning approaches have some problems. The most notable problem is that they require a large number of labeled training documents for accurate learning. While unlabeled documents are easily collected and plentiful, labeled documents are difficultly generated because a labeling task must be done by human developers. In this paper, we propose a new text classification method based on unsupervised or semi-supervised learning. The proposed method launches text classification tasks with only unlabeled documents and the title word of each category for learning, and then it automatically learns text classifier by using bootstrapping and feature projection techniques. The results of experiments showed that the proposed method achieved reasonably useful performance compared to a supervised method. If the proposed method is used in a text classification task, building text classification systems will become significantly faster and less expensive.     

Relevance feedback and cross-language information retrieval

This paper presents a study of relevance feedback in a cross-language information retrieval environment. We have performed an experiment in which Portuguese speakers are asked to judge the relevance of English documents; documents hand-translated to Portuguese and documents automatically translated to Portuguese. The goals of the experiment were to answer two questions (i) how well can native Portuguese searchers recognise relevant documents written in English, compared to documents that are hand translated and automatically translated to Portuguese; and (ii) what is the impact of misjudged documents on the performance improvement that can be achieved by relevance feedback. Surprisingly, the results show that machine translation is as effective as hand translation in aiding users to assess relevance in the experiment. In addition, the impact of misjudged documents on the performance of RF is overall just moderate, and varies greatly for different query topics.     

网络众包参与者创造力影响因素研究

本文基于工作投入和自我效能理论，对影响网络众包参与者创造力的关键路径进行分析，构建了网络众包参与者创造力影响因素模型。运用SEM模型对Kaggle网络众包竞赛平台的600位众包数据分析家的问卷调查数据进行实证检验，研究发现：（1）网络众包参与者的自我效能和工作投入都可以直接影响个人的创造力水平；（2）工作投入中的情感投入和认知投入对创造力有正向影响，而生理投入有负向影响；（3）自我效能中的领域任务自我效能和互动学习自我效能对创造力都有正向影响；（4）工作投入中的情感投入和认知投入在自我效能与创造力之间有中介作用；（5）自我效能中的领域任务自我效能和互动学习自我效能在工作投入与创造力之间也有中介作用。     

Factors affecting the selection of search tactics: Tasks,knowledge, process,and systems

This study investigated whether and how different factors in relation to task, user-perceived knowledge, search process, and system affect users’ search tactic selection. Thirty-one participants, representing the general public with their own tasks, were recruited for this study. Multiple methods were employed to collect data, including pre-questionnaire, verbal protocols, log analysis, diaries, and post-questionnaires. Statistical analysis revealed that seven factors were significantly associated with tactic selection. These factors consist of work task types, search task types, familiarity with topic, search skills, search session length, search phases, and system types. Moreover, the study also discovered, qualitatively, in what ways these factors influence the selection of search tactics. Based on the findings, the authors discuss practical implications for system design to support users’ application of multiple search tactics for each factor.     

Use of widgets as information management tools in online shared spaces

Online shared spaces are destinations where people of common interests interact to achieve common goals or borrow each other's expertise for personal pursuits. Due to a multitude of users, such spaces generate large amounts of content, thus creating a potential for information overload. To deal with it, users of online shared spaces develop information management strategies that frequently involve the use of information technology. In this paper we present a case study of the use of widgets as information management tools in uCern, an online shared workspace where healthcare industry stakeholders get involved in various types of interactions. To understand how widgets served the purpose of information management and whether they were considered useful, we administered an online questionnaire (n = 183) followed by qualitative interviews (n = 5). Data analysis revealed that study participants used widgets as tools that promoted their situation and activity awareness of events and activities going on in uCern and helped them with organizing and accessing the content in a quick and efficient manner. We also describe issues associated with the use of widgets and their usefulness as reported by participants and propose implications for design and development of online shared spaces in relation to information management tools.     

Ethical values and social care robots for older people: an international qualitative study

Values such as respect for autonomy, safety, enablement, independence, privacy and social connectedness should be reflected in the design of social robots. The same values should affect the process by which robots are introduced into the homes of older people to support independent living. These values may, however, be in tension. We explored what potential users thought about these values, and how the tensions between them could be resolved. With the help of partners in the ACCOMPANY project, 21 focus groups (123 participants) were convened in France, the Netherlands and the UK. These groups consisted of: (i) older people, (ii) informal carers and (iii) formal carers of older people. The participants were asked to discuss scenarios in which there is a conflict between older people and others over how a robot should be used, these conflicts reflecting tensions between values. Participants favoured compromise, persuasion and negotiation as a means of reaching agreement. Roles and related role-norms for the robot were thought relevant to resolving tensions, as were hypothetical agreements between users and robot-providers before the robot is introduced into the home. Participants’ understanding of each of the values—autonomy, safety, enablement, independence, privacy and social connectedness—is reported. Participants tended to agree that autonomy often has priority over the other values, with the exception in certain cases of safety. The second part of the paper discusses how the values could be incorporated into the design of social robots and operationalised in line with the views expressed by the participants.     

Search tactics used in solving everyday how-to technical tasks: Repertoire,selection and tenacity

With greater access to computational resources, people use search to address many everyday challenges in their lives, including solving technology problems. Although there are now many useful ‘how-to’ resources online (especially videos on YouTube), it can still be difficult to identify, understand, and resolve certain kinds of technical problem. While research tasks have been studied for many years and we know the tactics people use, we know far less about searchers’ tactics for how-to technical tasks that involve actually being able to apply found information to resolve a problem. Crucial to our study was developing and studying a highly realistic, how-to technical task, for which there was no single guidance resource: making a phone safe for a child. After providing 39 participants with an actual phone to fix, and a search engine to perform the task, we analysed their search tactics using retrospective cued think aloud interviews. Our primary contribution is a set of 77 tactics used, in three categories, along with detail of how common they were. We conclude that people had a lot of tactics in their repertoire. Although it was not hard for participants to find relevant information, what was hard was for participants to find information they could use; indeed only 23% of participants successfully completed the entire task. Domain knowledge affected the choice of tactics used (although not necessarily towards better task success). We discuss these influences and make design recommendations for how future search systems can support those in resolving how-to technical tasks.     

Local word vectors guiding keyphrase extraction

Automated keyphrase extraction is a fundamental textual information processing task concerned with the selection of representative phrases from a document that summarize its content. This work presents a novel unsupervised method for keyphrase extraction, whose main innovation is the use of local word embeddings (in particular GloVe vectors), i.e., embeddings trained from the single document under consideration. We argue that such local representation of words and keyphrases are able to accurately capture their semantics in the context of the document they are part of, and therefore can help in improving keyphrase extraction quality. Empirical results offer evidence that indeed local representations lead to better keyphrase extraction results compared to both embeddings trained on very large third corpora or larger corpora consisting of several documents of the same scientific field and to other state-of-the-art unsupervised keyphrase extraction methods.     

Large-scale evaluation framework for local influence theories in Twitter

Influence theories constitute formal models that identify those individuals that are able to affect and guide their peers through their activity. There is a large body of work on developing such theories, as they have important applications in viral marketing, recommendations, as well as information retrieval. Influence theories are typically evaluated through a manual process that cannot scale to data voluminous enough to draw safe, representative conclusions. To overcome this issue, we introduce in this paper a formalized framework for large-scale, automatic evaluation of topic-specific influence theories that are specialized in Twitter. Basically, it consists of five conjunctive conditions that are indicative of real influence exertion: the first three determine which influence theories are compatible with our framework, while the other two estimate their relative effectiveness. At the core of these two conditions lies a novel metric that assesses the aggregate sentiment of a group of users and allows for estimating how close the behavior of influencers is to that of the entire community. We put our framework into practice using a large-scale test-bed with real data from 75 Twitter communities. In order to select the theories that can be employed in our analysis, we introduce a generic, two-dimensional taxonomy that elucidates their functionality. With its help, we ended up with five established topic-specific theories that are applicable to our settings. The outcomes of our analysis reveal significant differences in their performance. To explain them, we introduce a novel methodology for delving into the internal dynamics of the groups of influencers they define. We use it to analyze the implications of the selected theories and, based on the resulting evidence, we propose a novel partition of influence theories in three major categories with divergent performance.     

Investigating the document structure as a source of evidence for multimedia fragment retrieval

Multimedia objects can be retrieved using their context that can be for instance the text surrounding them in documents. This text may be either near or far from the searched objects. Our goal in this paper is to study the impact, in term of effectiveness, of text position relatively to searched objects. The multimedia objects we consider are described in structured documents such as XML ones. The document structure is therefore exploited to provide this text position in documents. Although structural information has been shown to be an effective source of evidence in textual information retrieval, only a few works investigated its interest in multimedia retrieval. More precisely, the task we are interested in this paper is to retrieve multimedia fragments (i.e. XML elements having at least one multimedia object). Our general approach is built on two steps: we first retrieve XML elements containing multimedia objects, and we then explore the surrounding information to retrieve relevant multimedia fragments. In both cases, we study the impact of the surrounding information using the documents structure.     

略谈任务驱动法在工业设计人才培养中的应用

工业设计是一门具有较强实践性、创造性和时代性的专业,设计者必须通过设计作品来获得认可。任务驱动教学法以＂任务＂为核心,通过模拟真实生活情境、设计具有一定指向性的教学目标,激发起学生自主进行工业设计的内在动力,使学生在小组团结互助下,更好地完成学习任务,掌握工业设计的基础知识和技巧。在工业设计教学中,正确运用任务驱动法,使学生在有限的课堂教学环境中进行有效的实践创作,提高学生的设计实践能力,是工业设计教学改革的一条有效路径。     

How compliant are technicians with universal safety measures in medical laboratories in Croatia? – A pilot study

Introduction

This pilot study aimed to investigate the use of personal protective equipment (PPE) and compliance to the code of conduct (rules defined in institutional, governmental and professional guidelines) among laboratory technicians in Croatian medical laboratories. In addition, we explored the differences in compliance between participants of different age groups, laboratory ownership and accreditation status.

Materials and methods

An anonymous and voluntary survey with 15 questions was conducted among Croatian medical laboratory technicians (N = 217). The questions were divided into two groups: demographic characteristics and the use of PPE. The questions of the second part were graded according to the Likert scale (1-4) and an overall score, shown as median and range (min-max), was calculated for each participant. Differences between the overall scores were tested for each group of participants.

Results

The majority of participants always wear protective clothes at work, 38.7% of them always wear gloves in daily routine, more than 30.0% consume food and almost half of them drink beverages at workplace. A significantly lower overall score was found for participants working in public compared to private laboratories (36 (16-40) vs. 40 (31-40), P < 0.001). There were no statistically significant differences in overall scores for participants of different age groups (P = 0.456) and laboratory accreditation status (P = 0.081).

Conclusion

A considerable percentage of laboratory technicians in Croatian medical laboratories do not comply with safety measures. Lack of compliance is observed in all personnel regardless laboratory accreditation and participants’ age. However, those working in private laboratories adhere more to the code of conduct.Key words: survey, compliance, laboratory technician, universal safety measures, total quality management, clinical laboratory services