Ciao Roma! Glad to be attending #CoNEXT2022. Join us on Friday at our DistributedML workshop for some great talks and insights on the topic.

How do you scale training as models gets increasingly large and accurate? Great insights on programmable networks for performant distributed deep learning by @elfmar in the first session of DistributedML.

How can we make ML training and deployment at scale more efficient? Ana Klimovic' research sheds light on the data ingestion bottlenecks in ML processes.

We had a great time hosting the 3rd iteration of DistributedML. In case you missed it, we are delighted to release the video recordings of our workshop that you can find on distributedml.org/program and on YouTube youtube.com/@distributedml/. See you in the next one!

News is that DistributedML is going to #Paris this year with #CoNEXT23! Co-organised with Alexey Tumanov, Nathalie Baracaldo, Dimitrios Vytiniotis and Mario Almeida, @tforcworc. More information to follow soon ...

Reminder: Almost one week left until the submission deadline of distributedml.org (Sep 15th). We are looking forward to receiving some great papers this year! In the meantime, we have been organising an exciting program for this year's workshop. Stay tuned for more!

Pencils down in 11 days. Submission deadline for distributedml.org has now been extended for a week. New deadline is set for Sep 22nd AoE.

Announcing our speaker lineup for this year's DistributedML, including @ce__zhang (Together AI) , Tushar Krishna (Georgia Tech ECE ), @aurelien_bellet (@inria), Ada Gavrilovska (Georgia Tech School of Computer Science) and Dan Alistarh (MIT CSAIL and ISTAustria). Looking forward to a great program!

Great start of the day with @ce__zhang's insightful keynote on the current tensions of training and deploying LLMs.

Second keynote speaker of the day is Tushar Krishna, talking about hardware-software co-design in distributed systems.

We are now starting the third keynote of the day, with @aurelien_bellet presenting his work on Privacy Guarantees for Decentralised Federated Learning

The panel on the challenges and opportunities of the future of ML compute has started with @ce__zhang (Together AI) , Tushar Krishna (Georgia Tech ECE ), Ada Gavrilovska (Georgia Tech School of Computer Science), Dan Alistarh (MIT CSAIL and ISTAustria), Steve Laskaridis and Alexey Tumanov.

You can now find the recordings of this year's DistributedML on our website (distributedml.org/program/) and youtube channel (youtube.com/@distributedml). Thanks everyone for participating! See you in the next one.

Very excited to announce our TTODLer-FM workshop at ICML'25! Looking forward to some incredible contributions at the intersection of on-device ml and foundation models.