Data Conference Databasea Query.Farm index
Speakers

刘方鑫

Employers on the program

  1. 2026上海交大1 talk
Events1
Conferences1
Talks1
Years active2026

The year around now

Talks in the past 12 months and announced talks ahead, one row per conference, January to December. Hover a mark for the talk.

Held in the past 12 months Next 12 months Today, Nov 9, 2026

Appearances by year

201920202021202220232024202520262027
  • AICon 全球人工智能开发与应用大会 (AICon Global AI Development and Application Conference)

Talks

DateEventConferenceRoleTitleLinks
Jun 26, 2026AICon 全球人工智能开发与应用大会 | 上海AICon 全球人工智能开发与应用大会 (AICon Global AI Development and Application Conference)speaker从数据表征到数据流编排的存算协同优化
as 上海交大
当前大模型推理受限于非均匀数据分布与同构算力架构的严重错配,导致存储冗余、精度浪费与访存瓶颈。 针对上述问题,本报告提出一套面向大模型推理的跨层协同优化方案。在数据表征层面,通过分布感知的自适应数据编码,降低信息冗余度,实现模型参数的紧凑化与硬件友好型存储;在计算范式层面,重构运算逻辑,引入基于重要性感知的高精度近似计…Show less

当前大模型推理受限于非均匀数据分布与同构算力架构的严重错配,导致存储冗余、精度浪费与访存瓶颈。 针对上述问题,本报告提出一套面向大模型推理的跨层协同优化方案。在数据表征层面,通过分布感知的自适应数据编码,降低信息冗余度,实现模型参数的紧凑化与硬件友好型存储;在计算范式层面,重构运算逻辑,引入基于重要性感知的高精度近似计算,以低成本的轻量运算替代非关键数据的高精度运算,有效提升硬件算力利用率;在数据流层面,通过协同编排计算与访存数据流,优化调度策略,减少缓存未命中与流水线阻塞。为构建高效的 AI 算力底座提供了系统性的演进路径。 听众收益: ​学会诊断:从…

Frequent co-speakers

No one else spoke at two or more of the same events.

Speakers with the same name are merged into one entry, so this page may combine more than one person. Affiliations are shown as each conference listed them.