影响力指数

-/100

发表论文2 篇

平均评分3.0

年均产出2.0 篇/年

Soheil Zibakhsh Shabgahi

PhD student@University of California, San Diego·美国·OpenReview

研究方向

Learning based Optimization · Deep Learning · Deep Reinforcement Learning

MoEs Are Stronger than You Think: Hyper-Parallel Inference Scaling with RoE

ICLR 2026Rejected

MoE-PHDS: One MoE checkpoint for flexible runtime sparsity

ICLR 2026Rejected

合作者 (6)

Mohammad Samragh

Mehrdad Farajtabar