From Prof. Stefano Moretti, the Editor-in-Chief: Welcome to JMLFS
Dear colleagues,
Let me introduce you to the new Journal of Machine Learning in Fundamental Sciences (JMLFS).
The ML revolution is in full swing. In fact, the groundwork for it was prepared in the middle of the 20th century, yet, it is only with the ever continuing development of increasingly powerful computers, combined with computational algorithms refined over the past couple of decades, that the world has seen an explosion of applications of ML, in anything from health, to finance down to even autonomous cars!
However, just like any revolution, in order to succeed, the ML one not only needs to be perpetually reproducing itself while increasing in scope, which we have now witnessed over half a century, but it should also penetrate coherently all domains of its applicability, which it can only do if a common approach is developed across disciplines which cornerstone is, first and foremost, an assessment of its limitations. In essence, new applications of ML offering decision-making abilities require a full appreciation of uncertainty, which is necessarily embedded in research within the aforementioned fundamental sciences. Furthermore, the latter are now also predominantly computational and thus conversely provides a rich, open environment for the development of ML methods.
Hence, against this background, a need for a journal like JMLFS emerged. On the one hand, its vision is to foster a revolution, which is to accelerate the pace of scientific discovery in mathematics, physics, chemistry and biology, by offering a knowledge platform where to document the application of cutting-edge ML techniques to problems in any of these disciplines. On the other hand, armed with the full appreciation of scope and limitations in decision-making applications of ML that comes from a rigorous scientific background, it will aid the consequent disruptive change in the academic context as well as in sectors of the global economy.
Furthermore, for an approach that bears promise of such a dramatic change in our understanding of fundamental sciences and how the newly acquired knowledge can influence the world at large, it becomes mandatory to assess the ethical aspects of ML. That is, in parallel to driving this change, we ought to start recommending concepts of what is a right and wrong approach in the deployment of new ML tools, ie, in both building and using ML systems. While the connection to biology may be self-evident in a moral sense, this may seems more tenuous in relation to mathematics, physics, chemistry, yet, even in these contexts, we run the risk of generating theoretical biases that can affect such an understanding. Therefore, we aim at publishing papers covering these aspects too. Finally, with this in mind, in relation to biology and recognising its natural border with medical sciences, we will also consider papers wherein ML approaches are applied to the latter domain.
The journal aims at attracting top scientists as its contributors for the purpose of creating a multi-disciplinary pool of ML innovators adept at working in both academic and commercial environments. They will share their skills and visions through both journal submissions and a software repository. To our readers, we commit to publish only high quality, rigorously assessed research and tools which will impact either or both of these two domains. To our authors, we assert our commitment to a fair and impartial review process as well as unlimited accessibility of their outputs for the present and the future.
The creation of JMLFS could have not happened without the support of an outstanding international editorial board and the trust from Andromeda, its publisher. It has been a pleasure working with both of them in order to bring JMLFS to our readers and authors.