Lattice AI Research

Research focus

Architecture Design (Transformers, SSMs, MoE) (2)Distributed Systems & Hardware (2)Training Efficiency & Optimization (1)

Frequent co-authors

Deming Chen (1)Jason Cong (1)Azalia Mirhoseini (1)Christos Kozyrakis (1)

Papers (2)

Mar 5, 2026

Google Research6d ago·also BAIR, Meta AI, NVIDIA, Stanford HAI +8

AI+HW 2035: Shaping the Next Decade

Forget scaling compute – the future of AI hinges on a 1000x leap in energy efficiency via tight AI+Hardware co-design over the next decade.

Deming Chen, Jason Cong, Azalia Mirhoseini +34

Architecture Design (Transformers, SSMs, MoE)Distributed Systems & Hardware

6d ago

FlashAttention-4: Algorithm and Kernel Pipelining Co-Design for Asymmetric Hardware Scaling

FlashAttention-4 shatters attention bottlenecks on Blackwell GPUs, achieving up to 71% hardware utilization and 2.7x speedups over Triton, thanks to innovations like software-emulated softmax and asynchronous MMA pipelines.

Ted Zadouri, Markus Hoehnerbach, Jay Shah +4

Architecture Design (Transformers, SSMs, MoE)Distributed Systems & Hardware Training Efficiency & Optimization

Tri Dao

Publication activitypapers/week, last 8 weeks

Research focus

Frequent co-authors

Papers (2)

Search

Tri Dao

Publication activitypapers/week, last 8 weeks

Research focus

Frequent co-authors

Papers (2)