Lin Zheng
PulseAugur coverage of Lin Zheng — every cluster mentioning Lin Zheng across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
Scratchpad Patching improves byte-level language models
Researchers have introduced Scratchpad Patching (SP), a novel technique for byte-level language models that decouples compute requirements from patch size. This method addresses the trade-off between larger patches (red…
-
New framework improves federated learning for MoE LLMs on mobile devices
Researchers have developed FedAlign-MoE, a framework designed to improve federated learning for Mixture-of-Experts (MoE) large language models on mobile edge devices. This approach addresses challenges arising from data…
-
LLM long-context state tracking improved by deterministic compiler
A new approach to long-context state tracking in LLMs has been developed, demonstrating that traditional retrieval-augmented generation (RAG) methods are insufficient for maintaining derived state over extended document…
-
New GAPSL framework enhances federated learning in heterogeneous edge systems
Researchers have developed GAPSL, a new framework for parallel split learning designed to improve federated learning in edge computing systems with heterogeneous data. GAPSL addresses training divergence caused by incon…