The COVID-19 pandemic has accelerated many technological shifts that have already been taking place, including the shift from in-person to online and hybrid (in-person and online) academic gatherings. Eliminating the need for travel reduces environmental impacts and the burden imposed on delegates, and videotelephony services have demonstrated that they can meet the basic needs required to bring a gathering online. But while videotelephony can handle presentations and Q&A sessions, it cannot recreate the serendipitous hallway conversations that can lead to new scientific collaborations, ideas, and discoveries.   This grant supports an investigation by Roger Schonfeld at Ithaka Harbors, doing business as Ithaka S+R, into the opportunities and drawbacks of online and hybrid academic gatherings. Schonfeld and his team will use a combination of research, interviews, and design thinking—a human-centered approach to innovation—to help approximately 15 academic societies plan better academic gatherings that are informed by what works well in online and hybrid gatherings. The team’s project will elucidate the new skills and technologies needed to foster the interactions that were easy to create with in-person gatherings, which they will then share with decision-makers to enable more sustainable, inclusive, and engaging options for scholarly communication.

Innovation in and adoption of open hardware practices for scientific instrumentation and apparatus are being held back by the lack of widely-accepted standards in the description and versioning of open hardware projects. Metadata standards, in particular, are essential infrastructure to enable discovery and collaboration. A typical open source hardware project can combine instructions for 3D-printed components to be built locally along with a heterogeneous range of premade components (with different degrees of quality control) from a number of suppliers, along with any number of software programs used to control the device. At the moment, much open source hardware is in the “you can find documentation on my website” stage of maturity, where documentation and assembly instructions are idiosyncratic to the individual creator, and collaboration beyond small, local teams is more or less impossible.This grant funds Andrew Lamb, the founder of the Internet of Production Alliance, in a project to establish five families of metadata standards for open hardware: Designs and Documentation; Machines and Tools; People and Skills; Materials and Components; and Contracts and Business Models. These five standards are at different stages of maturity and build on each other: the first two (Open Know-How and Open Know-Where) have already been developed and activity will primarily focus on broader adoption and maintenance; the next two will be actively developed and launched over the course of the two years; and the last will be scoped for future development.

The Center for Scientific Collaboration and Community Engagement (CSCCE) has quickly become the preeminent research and training center focused on promoting the essential role community managers play in the effective functioning of scientific communities and thus in the production and dissemination of scientific knowledge. Led by microbiologist Lou Woodley, CSCCE documents and disseminates best practices in scientific community management, designs online and in-person curricula, runs training seminars, and acts as an advocate among scientists for the professionalization and institutionalization of the community management role. Funds from this grant support the continued operation and expansion of the CSCCE, along with efforts to develop and implement a business sustainability plan that will allow the organization to continue providing services to the diverse community of an estimated 30,000 community managers inside STEM research organizations. Grant funds are being administered by Community Initiatives, Inc., acting as a fiscal sponsor for CSCCE.

This grant funds the continued development of the AI Model Share platform, a website and integrated open-source Python library where researchers can deploy and share versions of machine learning models they have created in their research, and which can then be subsequently downloaded, implemented, used, analyzed, and improved by other researchers. In addition to making new resources available to researchers of all kinds, AI Model Share’s careful attention to issues like requirements tracking, versions, and documentation is an important step towards creating standards, tools, and practices that will allow research using machine learning methods to be robustly replicated. Activities supported by grant funds include the beta launch of the platform, user training and feedback workshops, an expansion of the platform’s ability to submit, search for, and replicate stored AI models, and the development of a front end “portfolio page” interface for platform users.