An official website of the United States government
A tool that could suggest new personalized research directions and ideas by taking insights from the scientific literature could significantly accelerate the progress of science. A field that might benefit from such an approach is artificial intelligence (AI) research, where the number of scientific publications has been growing exponentially over the last years, making it challenging for human researchers to keep track of the progress. Here, we use AI techniques to predict the future research directions of AI itself. We develop a new graph-based benchmark based on real-world data -- the Science4Cast benchmark, which aims to predict the future state of an evolving semantic network of AI. For that, we use more than 100,000 research papers and build up a knowledge network with more than 64,000 concept nodes. We then present ten diverse methods to tackle this task, ranging from pure statistical to pure learning methods. Surprisingly, the most powerful methods use a carefully curated set of network features, rather than an end-to-end AI approach. It indicates a great potential that can be unleashed for purely ML approaches without human knowledge. Ultimately, better predictions of new future research directions will be a crucial component of more advanced research suggestion tools.
more »
« less
Automatic Generation of Cancer Research Abstracts, AI vs Journal Articles
Introduction: There is an overwhelming amount of journal articles for modern researchers to parse through. For instance, there have already been 168,168 cancer-related papers archived on PubMed this year. In order to keep up with this substantial amount of literature, there are emerging interests in applying artificial intelligence (AI) to facilitate paper reading and drafting of new scientific ideas. Here, we extend the application of the state-of-the-art automatic research assistants to the cancer field. Using training datasets composed of over 5,000 cancer-related journal papers abstracts, we evaluated AI-based background knowledge extraction and abstract writing. The best AI performance is rated to be on par with human writers through a survey to university cancer researchers. This automatic research assistant tool can potentially speed up scientific discovery and production by helping researchers to efficiently read existing papers, create new ideas and write up new discoveries.
more »
« less
- Award ID(s):
- 1757885
- PAR ID:
- 10138560
- Date Published:
- Journal Name:
- 2019 BMES Conference Proceedings - REU Abstract Accepted Poster
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Given the inherent visual affordances of Head-Mounted Displays (HMDs) used for Virtual and Augmented Reality (VR/AR), they have been actively used over many years as assistive and therapeutic devices for the people who are visually impaired. In this paper, we report on a scoping review of literature describing the use of HMDs in these areas. Our high-level objectives included detailed reviews and quantitative analyses of the literature, and the development of insights related to emerging trends and future research directions. Our review began with a pool of 1251 papers collected through a variety of mechanisms. Through a structured screening process, we identified 61 English research papers employing HMDs to enhance the visual sense of people with visual impairments for more detailed analyses. Our analyses reveal that there is an increasing amount of HMD-based research on visual assistance and therapy, and there are trends in the approaches associated with the research objectives. For example, AR is most often used for visual assistive purposes, whereas VR is used for therapeutic purposes. We report on eight existing survey papers, and present detailed analyses of the 61 research papers, looking at the mitigation objectives of the researchers (assistive versus therapeutic), the approaches used, the types of HMDs, the targeted visual conditions, and the inclusion of user studies. In addition to our detailed reviews and analyses of the various characteristics, we present observations related to apparent emerging trends and future research directions.more » « less
-
Abstract Because research in science, engineering and medical fields advances incrementally, researchers routinely write papers that build directly on their prior work. While each new research article is expected to make a novel contribution, researchers often need to repeat some material—method details, background and so on—from their previous articles, a practice called ‘text recycling’. While increasing awareness of text recycling has led to the proliferation of policies, journal editorials and scholarly articles addressing the practice, these documents tend to employ inconsistent terminology—using different terms to name the same key ideas and, even more problematic, using the same terms with different meanings. These inconsistencies make it difficult for readers to know precisely how the ideas or expectations articulated in one document relate to those of others. This paper first clarifies the problems with current terminology, showing how key terms are used inconsistently across publisher policies for authors, guidelines for editors and textbooks on research ethics. It then offers a new taxonomy of text‐recycling practices with terms designed to align with the acceptability of these practices in common research writing and publishing contexts.more » « less
-
null (Ed.)In many academic fields, the number of papers published each year has increased significantly over time. Policy measures aim to increase the quantity of scientists, research funding, and scientific output, which is measured by the number of papers produced. These quantitative metrics determine the career trajectories of scholars and evaluations of academic departments, institutions, and nations. Whether and how these increases in the numbers of scientists and papers translate into advances in knowledge is unclear, however. Here, we first lay out a theoretical argument for why too many papers published each year in a field can lead to stagnation rather than advance. The deluge of new papers may deprive reviewers and readers the cognitive slack required to fully recognize and understand novel ideas. Competition among many new ideas may prevent the gradual accumulation of focused attention on a promising new idea. Then, we show data supporting the predictions of this theory. When the number of papers published per year in a scientific field grows large, citations flow disproportionately to already well-cited papers; the list of most-cited papers ossifies; new papers are unlikely to ever become highly cited, and when they do, it is not through a gradual, cumulative process of attention gathering; and newly published papers become unlikely to disrupt existing work. These findings suggest that the progress of large scientific fields may be slowed, trapped in existing canon. Policy measures shifting how scientific work is produced, disseminated, consumed, and rewarded may be called for to push fields into new, more fertile areas of study.more » « less
-
Abstract Translational research (TR) represents a promising systematic process for going from scientific discoveries to practical applications. Through conversations with academics, practitioners, decision‐makers and users, there has emerged a broad level of water science community support for including TR in Water Resources Research (WRR) publications. Based on this, we now open a continuing special collection of TR papers in WRR. The aim is to facilitate a community within hydrology and water science that seeks to provide actionable knowledge for societal benefit across disciplines, scales and contexts, with a focus on water as a key societal resource or a risk (e.g., of floods, droughts, or as pollutant carrier). This Editorial discusses what the multi‐faceted nature of TR may include in the context of WRR, why it is important to encourage TR papers in WRR, and how the opening of a continuing special collection of translational water research papers initiates a process to include such articles in the journal.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
