The Conference on the Mathematical Theory of Deep Learning (DeepMath) is a unique annual meeting dedicated to a truly cross-disciplinary approach to a fundamental understanding of the deep learning methodologies that are revolutionizing machine learning both in academics and in industry. This grant aims to broaden participation at DeepMath; a meeting where diverse researchers across backgrounds and disciplines come together to develop a fundamental theory of Deep Learning, and by proxy other large, distributed learning systems. The central philosophy mirrors how neuroscience began with a mix of physicists, biologists, psychologists, engineers, and mathematicians all interested in the same quandary. By definition, the foundation of such an effort is in broadening participation and maximizing diversity and inclusion to create a truly egalitarian effort.
This grant funds new programs aimed at ensuring that our community is accessible and welcoming to those from under-represented groups through (1) fellowship awards, (2) speaker costs (travel and lodging), and (3) broadcasting live streaming and video hosting to broader research community. Specific programs include outreach via targeted advertising and inclusion of members of these groups in the organizing committee, travel grants, childcare grants for economically disadvantaged researchers, and plans for a future mentorship programming collaboration with LatinXinAI and other professional organizations. Assessment via questionnaires and demographic data will determine which efforts succeeded most to continually refine these programs for future years.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
