We are now accepting abstract submissions to the 2020 Conference on the mathematical theory of deep neural networks (DeepMath). DeepMath is a highly interdisciplinary conference focused on understand fundamental theory driving the empirical success of Deep Learning. We are excited for this year's diverse set of invited speakers (see https://deepmath-conference.com for more information) and invite additional contributions from the community.

Due to ongoing concerns about Covid-19, this year’s DeepMath conference will be conducted virtually. We are presently working on finalizing an online format that will seek to capture the small, community-focused experience of past DeepMath events. Important updates related to submission and registration are included below: 

== Submissions == 
We are still accepting submissions for virtual “posters” consisting of a pre-recorded talk and Q&A session. Similarly to last year, a selected number of top posters will be invited to present their work live as contributed talks. 

== Registration == 
Participation in the conference will require registration costing a nominal fee. We are working to keep costs as low as we can given the online nature of the event. As with last year, all talks will be live-streamed for free.

More information on submissions can be found at the DeepMath website at
https://deepmath-conference.com/submissions

Warmest regards,
The DeepMath Organizing Committee