Efficient Systems for Foundation Models

Code it, run it, crash it–restart it.

➡️ ES-FoMO is back for ICML 2024! Check-out our 📝 call for papers, deadline 10th of June for submissions!

🔥 the gist

what? A workshop to bring together interdisciplinary experts working on the emerging research questions and challenges associated with foundation model training and inference.
when & where?
- Join us at ICML 2024, either virtually or physically in Vienna;
- Submit your papers on OpenReview (coming soon!), see the 📝 call for papers for more information.
questions? Contact us at esfomo.workshop@gmail.com.
looking for the 2023 edition?
- Check-out the recorded talks and panels, on the official ICML platform;
- Check-out the accepted papers on OpenReview.

📆 the plan

All times CET, UTC+2. Full schedule to be confirmed.

	Topic	Speaker
9:00am	Opening remarks
9:15am	📈 Session I: Emerging Architectures
	Mixture-of-Experts: Scaling and Tradeoffs at Training and Inference	Abhinav Venigalla (Databricks)
	A Deep Dive into State-Space Models	Albert Gu (Carnegie Mellon)
10:15am	🎤 Contributed Talk 1
10:30am	Coffee break
10:45am	🎤 Contributed Talk 2
11:00am	🚀 Session II: Efficient (and Open) Implementations
	Efficient Quantization Methods and Marlin, a Fast 4-Bit Inference Kernel	Elias Frantar (IST Austria)
	TBC
noon	Lunch break
1:00pm	🧑‍🎓 Poster Session
2:00pm	💬 Panel: Data and Architecture Trends Across Industry and Open Communities
	Aakanksha Chowdhery (Google DeepMind), Dylan Patel (SemiAnalysis), Stella Biderman (EleutherAI)–more to be announced!
	Moderators TBC.
3:00pm	🎤 Contributed Talk 3
3:15pm	Coffee break
3:30pm	🎤 Contributed Talk 2
3:45pm	⚙️ Session III: Data Tooling and Hardware
	Hardware for Efficient Machine Learning	Azalia Mirhoseini (Stanford)
	Open Tooling for Large Data Pipelines	Ludwig Schmidt (MIT/LAION)
4:45pm	🏅 Awards
6:00pm	🎉 Post-workshop happy hour

🦾 the pitch

As models increase in size and training budget, they not only systematically improve in upstream quality, but also exhibit novel emergent capabilities. This increase in scale raises proportionate difficulties for practitioners: foundation model training and inference lie at a unique interdisciplinary crossroad, combining open problems in algorithms, system design, and software engineering.

Machine learning practitioners are key stakeholders here: on the one hand, researchers may contribute algorithmic insights and novel methods to improving training and inference of large models; on the other hand, novel research findings may be best demonstrated at scale—which may require training models as efficiently as possible to make the best use of available resources.

The goal of this workshop is to bring together interdisciplinary experts working on the emerging research questions and challenges associated with foundation model training and inference. We welcome submissions around training and inference systems/algorithms for foundation models, focusing on scaling-up or on reducing compute, time, memory, bandwidth, and energy requirements. Notably, we encourage submissions concerning the entire spectrum of foundation models: from BERT-sized Transformers, to large models with 100B+ parameters. Topics include but are not limited to (see our 📝 call for papers for details):

Training and inference systems, either distributed at large scale or in resource-constrained scenarios;
Algorithms for improved training and inference efficiency;
Systems for foundation models, such as novel programming languages or compilers.

This is the second installment of ES-FoMo; we are bringing further focus in our sessions and talks on three trends observed in 2023:

The emergence of novel architectures, popularized by Mamba (state-space models) and Mixtral (mixture-of-experts);
Efficient open implementations, such as gpt-fast and vLLM;
Open questions on novel hardware and data tooling.