Lattice AI Research

Research focus

Architecture Design (Transformers, SSMs, MoE) (2)Distributed Systems & Hardware (2)Inference & Quantization (2)

Frequent co-authors

Xuehai Zhou (2)Yiming Liu (1)Zhiguang Wang (1)Zhiwei Ke (1)

Papers (2)

Jul 14, 2026

1w ago

Realizable N:M Sparse Transformer Inference via Search-Kernel Co-Design

Achieving over 2.2x inference speedup in Vision Transformers while maintaining accuracy reveals the untapped potential of hardware-software co-design in optimizing model performance.

Yiming Liu, Wenqi Lou, Zhiguang Wang +4

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

Mar 2, 2026

Mar 2, 2026·also OPPO, SYSU

Hermes: A Unified High-Performance NTT Architecture with Hybrid Dataflow

Forget slow NTTs: Hermes' hybrid dataflow architecture delivers up to 13.6x faster throughput for homomorphic encryption than GPUs.

Hang Gu, Hang Gu, Teng Wang +9

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

Search

Wenqi Lou

Publication activitypapers/week, last 8 weeks

Research focus

Frequent co-authors

Papers (2)