Dave Kochalko interview – how technology could restore trust in science publishing

Speeding up the research process, according to ARTiFACTS, requires turning research communication on its head by tracking, recording and sharing findings in small fragments and in real-time even before publishing them as full journal articles

Bashing science publishers have gone mainstream. Today, science publishing and publishing practices are often described using a lexicon reflective of a lack of trust: predatory journal, publication bias, adverse incentives, peer review problems, rip-off publishing, and the alliterative science-stymying paywall. Add to this the ‘tug of war’ over Open Access and the potential effects of the ambitious initiative Plan S, and one has to ask, how can trust in science publishing be restored?

This issue was discussed at this year’s NFAIS conference, on 14th February 2019, by Dave Kochalko, Co-Founder and Chief Academic Officer at Cambridge, Massachusetts-based scholarly publishing start-up ARTiFACTS. In a podcast interview with SciencePOD, Kochallko talks about findings solutions to this lack of trust. He believes the answer lies in refocusing on the science and scientists themselves. Specifically, he acknowledges that published work in the form of research papers continues to be valuable and relevant. However, he sees too much emphasis still being placed on publishing in high-impact journals. Instead, he believes, the emphasis should be on whether the scientist and the research team are practising “good science.”

Kochalko defines this good scientific practice as sharing results and research that, “can be verified, reproduced, built upon by other researchers.” And, he adds, “doing so in a more timely fashion.” This means that instead of attributing parenthood to an idea at the stage of a research paper, the genesis of ideas should be recorded and attributed in a far more granular manner to much smaller chunks of findings, as soon as they are formulated in a lab book, for instance.

What ARTiFACTS does is to allow scientists to “do three fundamental, yet powerful things,”  Kochalko points out. First, to establish proof of existence, authorship and confirm the provenance of their work at any time. Second, to protect and manage the IP they created while facilitating knowledge sharing. Third, to receive valid breakproof attribution and credit assignment at any point for any type of research output.

In doing so, scientists can make their findings available throughout the entire research process and not just after a protracted publication process. This is facilitated by ARTiFACTS use of graph database and distributed ledger (or blockchain) technology. The result, according to Kochalko, will be a “deep historical archive of published and discovered findings” that is accessible to the broader scientific community.

Science publishers are coming around to this refocus on the scientist and the tools and technologies that can help make this happen. “We’ve been very pleased that the reaction from publishers has moved from scepticism to thoughtful curiosity,” says Kochalko. This might be taken as a sign that the industry has acknowledged that not only trust in science publishers but in science itself requires turning attention away from a journal brand to what Kochalko describes as the “creative discoveries and contributions of the scientist.”

 

 

Photo credit: Dave Kochalko

Start-ups Survey Debunks Myths about Science Publishing Innovative Ecosystem

Yvonne Campfens discusses some of the surprising findings of her start-ups study with Sabine Louët, SciencePOD CEO, in a podcast

The first study of its kind related to the start-up ecosystem in the science publishing industry brought a lot of discussions during the APE2019 conference, held on 16th January 2019, in Berlin, Germany. The author of this study is Yvonne Campfens, an independent science publishing consultant, based in Amsterdam, The Netherlands. Focusing on 120 start-ups— 92 of which are independent— the study debunked some of the myths around the innovative science publishing ecosystem.

Campfens was motivated to start this study as she felt there were a lot of myths surrounding start-ups in publishing. “I was really curious, because I had seen so many new names over the last decade, and I was wondering what had happened with all these new players,” she says. “At the same time there was a lot of debate about distribution, developments, and trends in the industry,” she adds, “but very little facts and data.” This is what motivated Campfens to take a closer look.

She managed to debunk some of the myths surrounding current innovative companies in the field. “When a start-up is taken over, there is a lot of media attention, which gives some people the impression that this is all there is. So, I wanted to know what was really going on.” Specifically, she examined whether these start-ups existed independently from larger publishers.

First, the study reveals that there is now a well-established community of independent publishing start-ups providing innovation in the processes of publishing and content creation for the scholarly communication industry.  Their focus ranges from providing tools for discovery of information in research to offering solutions to prepare and analyse data from research. In addition, a number of start-ups deliver technology-based solutions to assist in writing about research. SciencePOD, which took part in the study, fits in the latter category.

Campfens also delved into the details of how and where their funding came from. She found that 10 start-ups raised $25 million or more in funding, another 7 raised $10 million or more while 29 companies raised $1 million or more. (See bar graphs below for a visual.)

 

Finally, she analysed when, during their lifecycle, start-ups were acquired and by whom. Of the 92 independent start-ups, she found that less than a quarter (21%) were acquired by other companies in 2018. It took, on average, 5 years of a start-up to be acquired. One of the most significant takeaways from her study, she says, “the acquired start-ups were definitely not taken over by STM players.” Indeed, the study reveals that companies that are not players in the science publishing industry acquired the majority (67%) of the 27 start-ups that were acquired.

To provide some first-hand comments from entrepreneurs themselves, Campfens invited a panel of start-up representatives. These included Arne Smolders, CEO of Academiclabs.co, based in Gent, Belgium.  This company is a social media network for linking academic and industry researchers in the life science sector.

Another panellist, Niklas Dorn, is CEO of video streaming solution Filestage.io based in Stuttgart, Germany, explained how scientific meetings could benefit from his solution.

Then, systems architect Roman Gurinovich with SCI.ai based in Tallinn, Estonia, outlined the benefits of his AI tool to validate scientific hypotheses from the published literature.

Last but not least,  Sami Benchekroun CEO and Co-founder of  Morressier based in Berlin, Germany, explained the concept of his publishing platform where researchers can share and showcase their early work at pre-publication stage, with the aim of fostering greater collaboration among scientists at an early stage.

If you liked this, please remember to like, share, and subscribe. Thank you very much and we hope to see you next time.

 

Andrew Burgess speaks about the use of AI in the content creation sector at ConTech

“The conversation is the relationship”, according to Management Consultant Andrew Burgess, who provides strategic insight around AI. In an interview recorded at ConTech 2018, held at the end of November in London, UK, Burgess talks with SciencePOD about the potential of AI in the content creation sector.

Should professionals in science publishing involved in creating content be concerned about their jobs as AI solutions become more prevalent?

“It will impact jobs but,[…] it’s going to be slower than people think. The things AI can do today is exciting and valuable, but there are limitations.” These limitations prevent AI from replacing people, entirely.

Rather they could augment people’s capabilities. Burgess continues: “AI’s very good at providing additional information to people, extracting value from data that humans wouldn’t necessarily be able to do, to help them make better decisions around their business… These are giving people additional capabilities that they wouldn’t have had before.”

What type of applications could help professionals involved in creating content augment their capabilities?

For instance, “in translation services, we’ve gone from a very structured approach to AI within the natural language to actually use machine learning, whereas the name suggests, it’s the machines that are learning how to do the translation based on the data that you’re giving them, ” Burgess notes. “And that gives you so much wider capabilities than you’ve ever had before… That’s where the real acceleration will be with the development of AI over the next 5-10 years.”

There are still concerns that AI may become harmful to humans? Where is the threat?

“I think the biggest risk is the people that are building and using the AI itself, all of this is just a tool that we use. As long as we plan right and don’t have a bias in the data, we don’t have opacity in the decisions the AI is making… (these) are really important things. That’s really up to humans to control.”

Enjoy the podcast and thanks for listening!

Podcast, Ivan Oransky, RetractionWatch: Retractions are like red flags highlighting infractions in science

Photo credit: SciencePOD.

 

Keeping a record of the retraction of research publications made it easier for journalistic coverage dissecting the infractions occurring within science publishing. “What I see that’s important, is not just coverage of individual cases, people are actually trying to put this all together. They’re filing public record requests because that’s something we’re not thinking of.” That’s according to RetractionWatch, co-founder, Ivan Oransky,  who started this initiative as a blog in 2010 with a single purpose in mind: making the peer-review process more transparent.

Listen to the edited version of the recording by SciencePOD.org of a recent presentation Oransky made at Trinity College Dublin, Ireland. This event, held on 20th June 2018, was co-organised by the Irish Science & Technology Journalists’ Association and the Science Gallery Dublin.

Oransky recalls the motivation that originally animated him and co-founder Adam Marcus in highlighting the mishaps of the peer-review process within academic communities. “Those of you who may be familiar with PubMed, Medline or Web of Science, you go to any of those you’ll find under 6,000 (retractions)… we [at Retraction Watch] have 3 times as many,” notes Oransky. Today, the RetractionDataBase.org site holds 17,500 retractions –and it is still growing. While retractions are rare, Oransky believes there is a screening effect attached to them.

For a sense of scale, the two countries in the world with the most retractions are China and the US. To provide an in-depth look at this, Oransky and his team compiled a leaderboard. Each of these instances are linked with a comprehensive story following the original publication.

Many varieties of malpractice

Oranksy highlights a few of the problems surrounding retractions found in the peer-review community. At the time of this recording, RetractionWatch had cataloged, 630 retractions specifically due to e-mail fraud by submitting a fake peer-reviewer’s e-mail. How does this work? An academic submits a paper to a journal for submission. When the journal comes back to ask for an e-mail to reference for peer-review, rather than submitting a genuine e-mail, the academic offers a fake e-mail, which then closes the loop between him or herself and the journal. Thus, eliminating the need for a peer-review. Back in 2000, only about 5% of papers were retracted, due to e-mail fraud.

Another area of malpractice occurs through duplication of results in different journals, not to be confused with plagiarism. Duplication is giving undue weight to a scientific conversation within the literature. This means when you try to conduct a scientific analysis on a topic, you’re looking at publications publishing the same thing multiple times without adding value to the topic.

All knowledge is provisional

To assume a paper should be retracted because the results aren’t reproducible is odd; but, it does occur. This shows that there is no perfect system for scholarly publishing. And that keeping a tap on retractions can help to uncover unsavoury behaviour among scientists.

Ultimately, this red flag activity leads to stronger science, as researchers are aware of the potential downsides of naming and shaming authors of retracted papers.

Enjoy the podcast!

 

Photo credit: SciencePOD.org

Podcast: How open science could benefit from blockchain

Experts reflect on the  implications of blockchain for research.

Reporting from the APE2018 a recent conference gathering the who’s who of scholarly publishing in Berlin on 16th and 17th January 2018, EuroScientist Editor, Sabine Louët interviews several experts on their views on how blockchain technology will change the world of scientists.

First, we hear from Lambert Hellerwho is the head of the Open Science Lab at TIB Hannover, the German national library for science and technology,  who gives his perspective as a digital librarian. He gives the bigger picture of how blockchain is going to help science become more open and help remove the current bottlenecks in the scientific endeavour by increasing the connectivity, accessibility and storage of scholarly objects, such as research papers and databases, through metadata and interplanetary data systems

Second, Amsterdam-based Joris van Rossum, director special projects, Digital Science, London, UK, highlights key findings of a recently published report about Blockchain for Research he has written. In particular, he outlines the many aspects of the research endeavour that could benefit from tracking, including through the use of blockchain technology, which can be in the form of data layer underneath the current research ecosystem.

Then, comesBerlin-based, Sönke Bartling,  founder of Blockchain for Science, whose mission is ‘to Open up Science and knowledge creation by means of the blockchain (r)evolution’ speaks about how blockchain could change the way science is being funded via the creation of cryptocurrencies.

Finally, we hear from  Eveline Klumpers, co-founder of Katalysis,  a start-up aiming to redefine the value of online content and focusing on developing blockchain solutions for the publishing industry. Based in Amsterdam, The Netherlands. She gives some concrete examples of how blockchain technology–which can store transparent immutable smart contracts defining how the content should be shared. This approach can give power back to authors of original work and help them monetise it. It could also help ensure reproducibility in research.

 

Original article published on EuroScientist.com.