AI summarisation to impact research productivity
Sabine Louet believes bite-size research can now be accessible anywhere
As the fuss over GPT-style technology continues, scientists are now fully aware that the future of research productivity will depend on Artificial Intelligence (AI). However, a different kind of AI, called extractive summarisation and based on natural language processing, will bring a much-needed time-saving tool to authors. Such a tool preserves the integrity of the evidence without ‘hallucinations’.
As SciencePOD announces the Widget version of its ScioWirebeta knowledge newsfeed, Sabine Louët, Founder and CEO, talks to Research Information about this innovative tool and how it creates structured research summaries, saving time in literature reviews, monitoring and reference searches. Such an AI summarisation tool provides instant ‘augmented insights’ of the content of publications at scale and delivers these summaries to a highly targeted audience through personalised newsfeeds.
Thus, by showcasing their content in a more in-depth manner than abstracts, ScioWire helps to make journals more attractive for submission, while increasing usage of their full-text studies. Ultimately, this approach is designed to boost publication impact.
SciencePOD has been at the forefront of providing specialist digital content based on scientific and clinical research for nearly nine years. Could you explain what ScioWirebeta is about?
ScioWirebeta is a cutting-edge system that uses AI to generate research newsfeeds. These feeds are made up of structured summaries of the latest scientific publications. The system allows researchers and professionals to create their own custom feed, tailored to their preferred keywords.
Our solution employs sophisticated natural language processing algorithms to turn complex studies into bite-sized briefings, in seconds. The summaries make it easy for users to quickly extract and grasp the essential points in a research study. They help scientists decide whether to dive further into the full text of the original study.
What makes these summaries so unique?
Summaries are provided with a contextual layer: providing main keywords, elucidated acronyms and lay definitions of technical terms. They also consist of a series of highlights (answering key questions about a study, such as Who? Where? When? What? How? What’s next?) together with an extractive summary. Extractive summarisation maintains the integrity of the research evidence, as it directly lifts sentences from a source study. This prevents the occurrence of ‘hallucinations’ associated with machine-learning driven tools such as GPT-4.
So, what does the ScioWire Widget bring to the existing ScioWire newsfeed?
The ScioWirebeta Widget can be embedded into a third-party web site, such as portals hosting communities of researchers, science conference microsites or closed websites reserved for healthcare professionals. It, therefore, brings custom newsfeeds directly to the professionals who need and work with research studies right where they already operate online.
What customisation options are available to individual users?
The ScioWirebeta Widget is highly customisable. Users can easily create their own feed, using filters such as keywords, journal title or publisher. Users can also choose to read the feed in their preferred language via automated translation. Once they have created their feeds, users access individual summaries of interest in a click.
Thus, it helps individual researchers or clinicians, as well as consultants, funders, or students to monitor the latest research, in their chosen topics. This solution saves time and increases efficiency in accessing the latest Open Access research.
How can innovative organisations, such as pharma companies or research institutes, benefit from the Widget?
The Widget can also be hosted on an innovative organisation’s chosen website. The Widget is designed to enhance user experience and boost audience engagement.
Several types of personalisation options are available to organisations. One of the main customisations is the ability to select a collection of carefully curated studies to be summarised and fed to the Widget. Pre-set filters then enable it to generate newsfeeds in the organisation’s chosen topics. They can also opt to let their own audience create their own newsfeeds. Further customisation stems from the addition of the organisation’s own logo on the Widget for enhanced brand recognition.
What do you see as the future of research summarisation and the role of AI in this domain?
The future of research summarisation is undoubtedly going to be shaped by advances in natural language processing and machine learning. As technology evolves, we'll see more refined and nuanced summarisation solutions. These not only save time but also help researchers make connections between different studies and discover new insights. AI will play a significant role in enhancing the research landscape, fostering greater accessibility, discoverability, and impact for scientific discoveries.
More information available at: www.sciowire.com