A couple of months ago, the scientific world was mesmerized by the tumultuous pour-down of new papers from the Artificial Intelligence research community. Studies diving into the capabilities of the hugely successful commercial ChatGPT, technical reports for aspiring open-source Large Language Model projects from university teams and tech giants, an undeniable gold-rush situation unfolded within the span of a few weeks. The flow has slowed-down since, but is still on-going.
Before that, the COVID outbreak produced a huge interest in scientific studies of epidemiology, vaccine production and efficacy.
All these experiences prompted me to ponder fundamentally about what assists a modern researcher’s daily struggle to follow the trends or communicate with colleagues, what is working about it and what is missing.
Two important factors that I figure make a researcher’s life better are:
To make a long story short, a researcher needs to follow-up on new research, and needs to be able to quickly spin-up a body of past literature to review. More often than not, they need to also be hit with something outside their bubble, to broaden their range and spark their inspiration. Second, in the information age they already need to be collaborating with a large number of associates, up or down the hierarchy.
Naturally, these two are communicating vessels, as fellow collaborators can assist in finding relevant research, and by contacting authors of new research you can reach a wider audience of scientists to work with.
I cannot refrain from recognizing that the above is connected to the more enveloping issues of accessibility to opportunity, funding and equipment, blocking points that all less-developed research communities and aspiring researchers struggle with today.
The issue going hand-in-hand with ease of discovery is that of accessibility. As it currently stands, most peer-reviewed publications are indeed behind a “pay-wall”, requiring either one-time purchase fees, monthly or yearly personal subscriptions, or membership in a subscribing public or private research institution.
Online social media applications have penetrated every aspect of society, and it is no wonder that they are also a place where discussion about research happens. Social media offer ways of interacting publicly, and also of creating private groups, thus allowing for any kind of conversation scope. The profile of social media applications ranges from relaxed (think Reddit) to more sometime more preppy (think LinkedIn), and research is indeed discussed under a wide spectrum of seriousness. Some allow their users to remain anonymous (or pseudonymous), and some encourage using your actual given name.
But as has been observed time and time again, an emergent property of social media is that they will be associated with a certain type of behaviour or content, and it seems hard to me to isolate this branding of existing platforms, and to re-purpose them, for example, as the place of a rather serious research discussion venture.
In practice, however, most researchers will publicly network within those platforms, mostly refraining from engaging in substantial public discourse, and perhaps create closed communities. Those communities, I imagine, are rather used more for generic coordination than for fine-grained exchange of ideas and brainstorming under specific research contexts. I’ve come to realize that such exchanges are mostly done in private using instant messaging apps, with various links to publication resources thrown around and a continuous unstructured stream of comments flowing through, until someone interrupts the flow to use the app for something more urgent that actually needs to instantly catch the group’s attention.
Regarding research hosting databases (such as PubMed or arXiv), most of them offer some sort of programmatic access (API), a means for developers in the wild to build code that connects to their archives and retrieves metadata, or even the actual research article text. But for pay-walled articles, the text not retrievable unless the application developed is in some way explicitly affiliated or subscribed to the journal source.
There’s another, perhaps not-so obvious reason why this is an important aspect of the research landscape, and relates also the ease of networking mentioned earlier. The reproduction of text from a pay-walled article is usually legally not allowed, therefore somewhat prohibiting detailed public discourse post-publication. This is why pre-prints are a thing, and this is why it’s hard to build an all-inclusive platform where all researchers can be un-restricted in elaborating their comments and critique.
The point is today’s researchers of all disciplines could benefit from tools that provide ease of discovery and ease of networking in research. Each of these tools has its own degree of focus to research content. One of my goals in this article is to make a list of the pros and cons of the most popular options I’m aware of, based on my understanding, as a starting point for a discussion. In the end, I’ll go through my own attempt at creating such a tool.
(Please feel free to contact me regarding factual errors in my review below. This does not claim to be an exhaustive list of features, pros & cons.)
Ups:
Downs:
Ups:
Downs:
Ups:
Downs:
Ups:
Downs:
Ups:
Downs:
Ups:
Downs:
Ups:
Downs:
Ups:
Downs:
Ups:
Downs:
Ups:
Downs:
Ups:
Downs:
Ups:
Downs:
Ups:
Downs:
After compiling this list, I have seen that there are many platforms out there to empower researchers (and I sure must have missed a lot of them). However, it is often the case that not all these tools are utilized to their maximum potential. I believe that mainstream attraction depends on:
Out of all the options presented to the researcher, I believe there are many satisfactory approaches to searching for publications. The options regarding discovery and recommendation of papers are still competing against one another, and frequently appear as paid services, and have also taken in the AI boom and seemingly integrated Machine Learning recommendation algorithms.
However, I feel that the networking and discussion aspect remains anchored in the general-purpose social media and messaging platforms. I have found discussion elements in some platforms. I can’t exactly verify the extent to which these are used in the global research community, but, for example, PubPeer has been around for around a decade, HuggingFace’s popularity has sky-rocketed and ResearchRabbit claims to be used by researchers in over 100 countries.
I personally had the idea of an alternative to social media for public and group discussions on user-curated paper collections before having explored the above app landscape, and it inspired me to create the Journal Hub platform. It shares many common elements with the options listed above, and encapsulates my vision of what such a platform should facilitate at minimum:
Outside this set many other attractions are found in the wild, such as visualization and personalized recommendation features. I haven’t used such tools during my research activity and I am wary of their effectiveness, but they may be helpful to others.
An all-in-one platform for research discovery and networking may be a goose chase, or it may be a matter of time.