3 papers
cs.LG2026
Mixture-of-Experts with Gradient Conflict-Driven Subspace Topology Pruning for Emergent Modularity
Yuxing Gan, Ziyu Lei
Mixture-of-Experts (MoE) architectures achieve parameter efficiency through conditional computation, yet contemporary designs suffer from two fundamental limitations: structural pa…
cs.DB2025
Is it Bigger than a Breadbox: Efficient Cardinality Estimation for Real World Workloads
Zixuan Yi, Sami Abu-el-Haija, Yawen Wang +8
DB engines produce efficient query execution plans by relying on cost models. Practical implementations estimate cardinality of queries using heuristics, with magic numbers tuned t…
cs.DB2025
Is Long Context All You Need? Leveraging LLM's Extended Context for NL2SQL
Yeounoh Chung, Gaurav T. Kakkar, Yu Gan +2
Large Language Models (LLMs) have demonstrated impressive capabilities across a range of natural language processing tasks. In particular, improvements in reasoning abilities and t…