Though multiple models of Open Access exist, the scientific community in the field of machine intelligence research does not seem to pay heed to the need of Open Access journals to demonstrate economic sustainability, to be viable without external support in the long term.
A Blog Article by Pablo Markin.
In early May, 2018, multiple news outlets have covered a scientific community outcry over the creation of a new closed-access journal in the domain of machine learning, recently launched by Nature journal as a topical area imprint. In a statement signed by over 2,500 supporters, Nature Machine Intelligence journal, a call is made to the members of the corresponding scientific community not to publish in this journal, review its submissions or participate in its activities as editors. The declared reason for this boycott of the fledgling journal that Tom Dietterich, its author from Oregon State University that also serves as a founding president of International Machine Learning Society, proposes is that only Open Access journals charging no article processing charges are welcome in this field of scientific inquiry.
On the one hand, Dietterich cites the resignation letter of one of editorial board members of the closed-access Machine Learning journal, as the rival Open Access Journal of Machine Learning Research was founded in 2001 to state that journals that serve scientific community needs ought to provide “immediate and universal access” to their articles. On the other hand, however, the citation also indicates that Open Access journals are expected to be doing that “at a cost that excludes no one,” which does not rule out APCs, such as those Gold Open Access journals have, or at-cost subscription charges, such as those of Green Open Access journals. Moreover, the majority of scientific journals in this scholarly field are both Open Access and do not involve APCs, such as the Conference and Workshop on Neural Information Processing Systems (NIPS), most probably indicates that they have external sources of financial support, especially given that machine intelligence is increasingly amenable to business applications.
Indeed, the NIPS conference has been publishing its proceedings from 1987, while its treasurer, Marian Stewart Bartlett, is associated with Apple Inc. and among its editorial and advisory board members there are representatives of Google, Google Research, IBM Watson Labs and Microsoft Research as well as those of universities from around the world. Similarly, the editorial and advisory board of the Open Access Journal of Machine Learning Research, which does not charge any APCs for its submissions, has both professors from globally leading universities, such as Harvard and Princeton, and numerous scholars from multinational corporations, such as Google, Facebook, Microsoft, Netflix, IBM, Baidu, and eBay. Thus, it appears that these Open Access journals that do not have APCs have close editorial and organizational ties to universities and corporations with multibillion endowments and revenues respectively.
Moreover, some commentators suggest that in the rapidly growing, and apparently lucrative and contested field of artificial intelligence Open Access journals play an important role in strengthening the collaboration between business and academia. For this reason, Springer Nature has been entering this market, where Open Access to empirical data and software tools has increasingly become the predominant norm.
By Pablo Markin
Featured Image Credits: Facebook VP of Engineering Regina Dugan F8, April 19, 2017 | © Courtesy of Anthony Quintano/Flickr.