In April, Meta launched the 2022 Network for AI request for proposals (RFP). Today, we’re announcing the winners of this award.
Every day, billions of people connect through Facebook, Instagram, WhatsApp, Messenger, and many other Meta products. Network infrastructure plays an important role in our business’s stability and scalability. Given the recent advances in AI computing, network infrastructure should adapt to meet the ever-changing needs of this new computing paradigm.
Through this RFP, we hope to support academics invested in new technologies to facilitate our next generation of AI applications.
The RFP attracted 38 proposals from 33 universities and institutions around the world. Thank you to everyone who took the time to submit, and congratulations to the winners.
Principal investigators are listed first unless otherwise noted.
Accelerating deep learning at the network interface on FPGA-DLA platforms
Mohamed Abdelfattah (Cornell University)
Accelerating distributed learning with compressed switch aggregation
Ran Ben Basat (University College London), Michael Mitzenmacher, Minlan Yu (Harvard University)
MILES: Multi-device incremental learning on edge via summarization
Yingyan Lin, Jyotikrishna Dass (Rice University)
Optimizing ML workloads via placement and collective communication
Danyang Zhuo, Matthew Lentz (Duke University)
SmartNIC-driven scheduling in AI/ML clusters
Mosharaf Chowdhury (University of Michigan)
Superscaling AI training with emerging electrical and optical interconnects
Manya Ghobadi (Massachusetts Institute of Technology)