AdaptFM

Program

Below is the full-day program of the workshop, taking place on July 11th in ASEM Ballroom 202.

Time (Local) Session Description
08:30 – 08:45 Welcome & Introduction Introductory session by the organizers on the topic, submissions, and scope of the venue.
08:45 – 09:30 Invited Keynote #1
When AI Starts Building AI		 Keynote talk by Beidi Chen.
09:30 – 10:00 Contributed Talks – Session 1 Oral presentations of contributed work.
· MAGE: All-[MASK] Block Already Knows Where to Look in Block Diffusion LLM
· Block-Level Recursion: Adaptive Test-Time Routing in Large Language Models
10:00 – 10:15 Coffee Break Coffee break and networking opportunities.
10:15 – 11:00 Invited Keynote #2
Accurate Low Bit Quantization Through Function Preserving Transformations		 Keynote talk by Markus Nagel.
11:00 – 11:30 Contributed Talks – Session 2 Oral presentations of contributed work.
· BASTION: Budget-Aware Speculative Decoding with Tree-structured Block Diffusion Drafting
· On State Reduction in Linear Attention
11:30 – 12:15 Invited Keynote #3
Coordinating Adaptive Inference: From Executable Model States to Agent Trajectories		 Keynote talk by Yingyan (Celine) Lin.
12:15 – 13:00 Lunch Break Lunch and networking opportunities.
13:00 – 13:45 Poster Session #1 Poster session of contributed work.
13:30 – 14:15 Invited Keynote #4
Towards Efficient Multimodal Intelligence		 Keynote talk by Hongxu (Danny) Yin.
14:15 – 14:30 Best Paper & Poster Awards Announcement of paper/competition awards.
14:30 – 15:15 Poster Session #2 Poster session of contributed work.
15:00 – 15:30 Coffee Break Coffee break and networking opportunities.
15:30 – 16:15 Invited Keynote #5
Adapting Foundation Models to Any Budget (Reuse & Retrieve, Don’t Retrain or Recompute)		 Keynote talk by Sangdoo Yun.
16:15 – 16:30 Contributed Talks – Session 3 Talks by top-3 competition teams (5 min each).
16:30 – 17:15 Panel Discussion Panel discussion with selected invited speakers.