Accelerating Persistent Neural Networks at Datacenter Scale

August 22, 2017
Eric Chung and Jeremy Fowers | Microsoft

Source: Hot Chips 29 (2017) Session 6

At 33:45 Eric Chung and Jeremy Fowers begin their presentation on Accelerating Persistent Neural Networks at Datacenter Scale

Projet
- Project Catapult

Regardez suivant

Dion2: A new simple method to shrink matrix in Muon
March 3, 2026
Anson Ho,

Kwangjun Ahn
ARO: A new lens on matrix optimization for LLMs
March 3, 2026
Anson Ho,

Wenbo Gong,

Chao Ma
Lessons from deploying HealthBots with experts-in-the-loop
March 3, 2026
Anson Ho,

Mohit Jain
Teaching small language models to think like optimization experts with OptiMind
March 3, 2026
Anson Ho,

Xinzhi Zhang
Agent Lightning: One learning system that makes all agents evolve
March 3, 2026
Anson Ho,

Luna K. Qiu
Magentic Marketplace: Testing societies of agents at scale
March 3, 2026
Gagan Bansal,

Anson Ho
CROSS — Leveraging AI ASICs for Homomorphic Encryption
February 26, 2026
Jianming Tong
Beyond Swahili: Designing Inclusive AI for Bantu Languages
February 17, 2026
Alfred Malengo Kondoro
Bridging Neurotechnology with Immersive Systems: Getting BCIs outside of the lab?
February 12, 2026
Hakim Si-Mohammed
Efficient Distributed Orthonormal Optimizers for Large-Scale Training
February 12, 2026
Kwangjun Ahn