Talks
Date Talk Venue / Host Location
Aug 2026 Efficient AI: From the Post-Processing Paradigm to Efficiency-Native AI ChinaMM 2026 Taiyuan
Jul 2026 Efficient AI: From the Post-Processing Paradigm to Efficiency-Native AI ChinaLLM 2026, CSIG Hangzhou
May 2026 The Two Paths to Efficient Inference for Multimodal LLMs: Exact Computing and Approximate Computing CCIG 2026 Guangzhou
May 2026 The Two Paths to Efficient Inference for Multimodal LLMs: Exact Computing and Approximate Computing VALSE 2026 Wuhan
Nov 2025 Model Redundancy vs. Data Redundancy: The Two Paths Towards Efficient MLLMs Based on Approximated Computing LMG 2025 Chengdu
Aug 2025 Recent Advances in Token Compression for Long-Context Multimodal LLMs VALSE Webinar Online
Jul 2025 LLM Compression and Acceleration at Low Cost ChinaLLM 2025, CSIG Hangzhou