This article provides the final results of a 3‐year study that sought to discover whether early career researchers (ECRs) were the harbingers of change with respect to scholarly communications. Over a hundred science and social science ECRs from seven countries, spanning three continents, were depth‐interviewed annually for 3 years (2016–2018) about their attitudes and behaviours with respect to 23 scholarly issues and activities (aspects). In order to provide an accessible overarching assessment of an extremely large and complex dataset, the interview data were categorized according to the strength and direction of change exhibited and the trends and points of interest raised. Results show that all ECRs have changed in one way or another, and a small minority has changed greatly in both attitude and practice and that collaboration and research impact are the scholarly aspects where most changes have occurred, and the greatest cause of change is not so much new technology as a change of jobs.

This study presents findings from the first year of the Harbingers research project, a 3-year longitudinal study of early career researchers (ECRs), which sought to ascertain current and changing habits in scholarly communication. The study recruited 116 science and social science ECRs from seven countries who were subject to in-depth interviews, and this paper reports on findings regarding publishing and authorship practices and attitudes. A major objective was to determine whether ECRs are taking the myriad opportunities proffered by new digital innovations, developing within the context of open science, open access, and social media, to publish their research. The main finding is that these opportunities are generally not taken because ECRs are constrained by convention and the precarious employment environment they inhabit and know what is bestfor them, which is to publish (in high impact factor journals) or perish.

Statistics over the most recently observed data elements are often required in applications involving data streams, such as intrusion detection in network monitoring, stock price prediction in financial markets, web log mining for access prediction, and user click stream mining for personalization. Among various statistics, computing quantile summary is probably most challenging because of its complexity. In this paper, we study the problem of continuously maintaining quantile summary of the most recently observed N elements over a stream so that quantile queries can be answered with a guaranteed precision of ǫN . We developed a space efficient algorithm for pre-defined N that requires only one scan of the input data stream and O( log(ǫ 2 N ) ǫ + 1 ǫ 2 ) space in the worst cases. We also developed an algorithm that maintains quantile summaries for most recent N elements so that quantile queries on any most recent n elements (n ≤ N ) can be answered with a guaranteed precision of ǫn. The worst case space requirement for this algorithm is only O( log 2 (ǫN ) ǫ 2). Our performance study indicated that not only the actual quantile estimation error is far below the guaranteed precision but the space requirement is also much less than the given theoretical bound.

This article presents findings from the first year of the Harbingers research project started in 2015. The project is a 3‐year longitudinal study of early career researchers (ECRs) to ascertain their current and changing habits with regard to information searching, use, sharing, and publication. The study recruited 116 researchers from seven countries (UK, USA, China, France, Malaysia, Poland, and Spain) and performed in‐depth interviews by telephone, Skype, or face‐to‐face to discover behaviours and opinions. This paper reports on findings regarding discovery and access to scholarly information. Findings confirm the universal popularity of Google/Google Scholar. Library platforms and web‐scale discovery services are largely unmentioned and unnoticed by this user community, although many ECRs pass through them unknowingly on the way to authenticated use of their other preferred sources, such as Web of Science. ECRs are conscious of the benefits of open access in delivering free access to papers. Social media are widely used as a source of discovering scholarly information. ResearchGate is popular and on the rise in all countries surveyed. Smartphones have become a regularly used platform on which to perform quick and occasional searches for scholarly information but are only rarely used for reading full text.

This paper presents selected findings from the first year of a 3‐year longitudinal study of early career researchers (ECRs), which sought to ascertain current and changing habits in scholarly communication. Specifically, the aims of the paper are to show: (1) how much experience and knowledge ECRs had of peer review – both as authors and as reviewers; (2) what they felt the benefits were and what suggestions they had for improvement; (3) what they thought of open peer review (OPR); and (4) who they felt should organize peer review. Data were obtained from 116 science and social science ECRs, most of whom had published and were subject to in‐depth interviews conducted face‐to‐face, via Skype, or over the telephone. An extensive literature review was also conducted to provide a context and supplementary data for the findings. The main findings were that: (1) most ECRS are well informed about peer review and generally like the experience, largely because of the learning experiences obtained; (2) they like blind double‐peer review, but would like some improvements, especially with regards to reviewer quality; (3) most are uncomfortable with the idea of OPR; and (4) most would like publishers to continue organizing peer review because of their perceived independence.

Early career researchers (ECRs) are of great interest because they are the new (and biggest) wave of researchers. They merit long and detailed investigation, and towards this end, this overarching paper provides a summary of the first‐year findings of a 3‐year, longitudinal study of 116 science and social science ECRs who have published nearly 1,200 papers and come from 7 countries and 81 universities. ECRs were interviewed in their own languages face‐to‐face, by Skype, or telephone. The study focused on the attitudes and behaviours of ECRs with respect to scholarly communications and the extent to which they are adopting new and disruptive technologies, such as social media, online communities, and Open Science. The main findings include: publishing in high‐impact factor journals is the only reputational game in town; online scholarly communities, and ResearchGate in particular, are gaining ground; social media are beginning to have an impact, especially in the dissemination arena; outreach activities have become more important; libraries are becoming increasingly invisible to ECRs; Open Science is not gaining traction; and more transformational ideas are being expressed, especially in the US and UK.

This article describes an international study informed by a 3-year-long qualitative longitudinal project, which sought to discover the scholarly communication attitudes and behaviour of early career researchers (ECRs). Using a combination of small-scale interviews and a larger-scale survey, ECRs were questioned on their searching and reading behaviour, publishing practices, open data, and their use of social media. Questionnaire invitations were sent out via publisher lists, social media networks, university research networks, and specialist ECR membership organizations. Onethousand and six-hundred responses were received, with many coming from China, Russia, and Poland. Results showed that ECRs are adopting millennial-facing tools/platforms, with Google, Google Scholar, social media, and smartphones becoming embedded in their scholarly activities. Open data sharing obtains widespread support but somewhat less practice. There are some differences in attitudes and behaviour according to age and subject specialism.

Recently, researchers have demonstrated that both Chinese word and its component characters provide rich semantic information when learning Chinese word embeddings. However, they ignored the semantic similarity across component characters in a word. In this paper, we learn the semantic contribution of characters to a word by exploiting the similarity between a word and its component characters with the semantic knowledge obtained from other languages. We propose a similaritybased method to learn Chinese word and character embeddings jointly. This method is also capable of disambiguating Chinese characters and distinguishing non-compositional Chinese words. Experiments on word similarity and text classification demonstrate the effectiveness of our method.